Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

Follow-up of breast cancer in primary care vs specialist care: results of an economic evaluation

Follow-up of breast cancer in primary care vs specialist care: results of an economic evaluation British Journal of Cancer (1999) 79(7/8), 1227–1233 © 1999 Cancer Research Campaign Article no. bjoc.1998.0197 Follow-up of breast cancer in primary care vs specialist care: results of an economic evaluation 1,2,6 4 1,5 1 1 6 7 3 1 E Grunfeld , A Gray , D Mant , P Yudkin , R Adewuyi-Dalton , D Coyle , D Cole , J Stewart , R Fitzpatrick and M Vessey 1 2 Division of Public Health and Primary Health Care, Institute of Health Sciences, University of Oxford, Oxford OX3 7LF, UK; Ottawa Regional Cancer Centre, Department of Medicine, University of Ottawa, Ontario K1H 8L6, Canada; Northamptonshire Centre for Oncology, Northampton General Hospital, Northampton 4 5 NN1 5BD, UK; Health Economics Research Centre, Institute of Health Sciences, University of Oxford, Oxford OX3 7LF, UK; Department of Primary Medical Care, University of Southampton, Southampton SO1 6ST, UK; Clinical Epidemiology Unit, Loeb Research Institute, University of Ottawa, Ontario K1Y 4E9, Canada; Princess Margaret Hospital, Swindon SN1 4JU, UK Summary A randomized controlled trial (RCT) comparing primary-care-centred follow-up of breast cancer patients with the current standard practice of specialist-centred follow-up showed no increase in delay in diagnosing recurrence, and no increase in anxiety or deterioration in health-related quality of life. An economic evaluation of the two schemes of follow-up was conducted concurrent with the RCT. Because the RCT found no difference in the primary clinical outcomes, a cost minimization analysis was conducted. Process measures of the quality of care such as frequency and length of visits were superior in primary care. Costs to patients and to the health service were lower in primary care. There was no difference in total costs of diagnostic tests, with particular tests being performed more frequently in primary care than in specialist care. Data are provided on the average frequency and length of visits, and frequency of diagnostic testing for breast cancer patients during the follow-up period. Keywords: breast cancer; economic evaluation; follow-up care; randomized controlled trial; primary care After completing primary treatment, it is standard practice in most 1996). We report here the results of the economic evaluation to countries for breast cancer patients to be followed in specialist assess the relative costs of the two alternative schemes of follow- outpatient clinics during the disease-free interval. This practice of up that was conducted concurrent with the RCT. Because the RCT post-treatment surveillance has recently come under close scrutiny showed no important differences in the primary clinical outcomes, (Dewar, 1995; Donegan, 1995; Loprinzi, 1995; Breast Cancer the form of economic evaluation was cost minimization whereby Surveillance Expert Panel, 1997). Studies to evaluate its effective- the two schemes are examined for the least costly alternative ness have used a Ôless intensiveÕ specialist follow-up regimen as (Drummond et al, 1997). the comparator (GIVIO Investigators, 1994; Rosselli Del Turco et al, 1994). No previous studies have evaluated prospectively the cost effectiveness of this practice of long-term follow-up of breast PARTICIPANTS AND METHODS cancer patients. Participants and methods for the RCT We conducted a randomized controlled trial (RCT) comparing specialist follow-up with follow-up by the patientÕs own general Participants were 296 women with breast cancer in remission practitioner (Grunfeld et al, 1996). The results showed no increase receiving regular follow-up at two district general hospitals in in delay in diagnosing recurrence and reinitiating specialist care as England. These women were taking part in an RCT to evaluate a a result of primary care follow-up. As has been reported in primary-care-based system of routine breast cancer follow-up, previous studies (Clark and Morris, 1981; Hughes, 1985; Tomin whereby they were randomized to one of two groups: continued and Donegan, 1987; McWhinney et al, 1990; Worster et al, 1995), routine follow-up in hospital outpatient clinics according to usual the results of the RCT (Grunfeld et al, 1996) showed that most practice (hospital group) or routine follow-up from their own recurrences are detected in the interval between regularly sched- general practitioner (general practice group). Ethical approval to uled follow-up appointments, and many are presented to the conduct the study was obtained from the local research ethics general practitioner irrespective of the formal follow-up arrange- committees. The recommended follow-up regimen was the same ments. The results also showed no increase in anxiety or deteriora- for women in both groups and involved periodic physician visits tion in health-related quality of life (HRQOL) (Grunfeld et al, for a breast cancer check-up, routine surveillance mammograms (the frequency depended on initial treatment and age), and diag- Received 18 May 1998 nostic testing only if clinically indicated. A full description of Accepted 30 July 1998 study participants, study methods and results of the primary outcomes of time to diagnosis of recurrence and HRQOL have Correspondence to: E Grunfeld, Ottawa Regional Cancer Centre, been reported previously (Grunfeld et al, 1996). Data collection Department of Medicine, University of Ottawa, 501 Symth Road, Ottawa, Ontario K1H 8L6, Canada for economic evaluation was an integral part of the RCT. 1227 1228 E Grunfeld et al Measurement and analysis of costs Analysis The two-tailed t-test was used to assess the significance of The perspective of the economic evaluation considered costs to the between group differences in means. The c test was used to assess health service (particularly, the costs of follow-up visits and diag- the significance of between group differences in proportions. nostic tests) and costs to the patients (such as lost earnings and out- Statistical analysis was performed using the software package of-pocket expenses). Valuation of resources was based on unit Statistical Package for the Social Sciences (SPSS, version 6.1.2). costs from the providers and national averages. All costs are Confidence intervals were calculated using the statistical program expressed in 1994 UK £s. Discounting was not considered to be CIA (Garner, 1991). All analysis was on an intention-to-treat relevant to this analysis, which concerns an 18-month time period. basis. Health service costs Information on health service resource use was collected prospec- Response rates tively at both hospital and general practice visits by means of a Completed record-of-visit forms were returned by physicians for record-of-visit form. This was completed at all consultations 100% (148 out of 148) of patients in the general practice group possibly related to breast cancer by doctors for patients in both (i.e. at least one form was received for each patient) and for 95.3% groups and provided details of the frequency of visits, the duration (141 out of 148) of patients in the hospital group. of each visit, and the diagnostic tests ordered at each visit. The The baseline cost questionnaire was completed by 99.3% (147 study hospitals provided unit costs for each type of test and for an out of 148) of patients in the general practice group and 95.3% average outpatient consultation (based on the total annual running (141 out of 148) of patients in the hospital group. By midtrial, four and capital costs of the department averaged across the total patients had died in the general practice group and five had died in number of attendances); the equivalent total costs for an average the hospital group. The adjusted response rate for the midtrial cost general practice consultation was taken from national averages. questionnaire was 97.2% (140 out of 144) and 88.7% (126 out of In the analysis of costs, the average of the unit costs for the 142) respectively. At the end of the study period, a total of six diagnostic tests, procedures and outpatient consultations provided patients had died in the general practice group and one had moved by the two study hospitals was used.* For estimates of the general and so was lost to follow-up, whereas a total of 12 patients had practitioner consultations, UK Department of Health estimates of died in the hospital group and one had moved and so was lost to the cost of a consultation were adjusted to allow for the differences follow-up. The adjusted response rate for the end of the trial cost in the average length of surgery and home consultations (Wilson, questionnaire was 97.2% (137 out of 141) in the general practice 1991). Using this approach, a cost per minute of £1.15 was esti- group and 88.1% (119 out of 135) in the hospital group. mated. To calculate a cost per GP consultation, the cost per minute As the adjusted response rate in the hospital group had fallen to was then applied to the length of each consultation. just above 88% while remaining above 97% in the general practice group, non-respondents in the hospital group were compared on Patient costs baseline characteristics and selected domains of the HRQOL A questionnaire to collect data on costs incurred by the patient was instruments. There were no significant differences between non- developed specifically for this study. The cost questionnaire respondents and respondents on any of these variables. related to costs associated with the patientÕs most recent breast cancer follow-up visit, and included questions about out-of-pocket expenses, lost earnings as a result of time off work, time spent in RESULTS attending the follow-up visit, and costs incurred by accompanying Health service resource use and costs persons. When patients, or those accompanying them, indicated that they had lost earnings over a specified period as a result of a During the 18 months of the study, general practice patients were follow-up visit, the cost to them of this visit was estimated by seen significantly more frequently (mean of 3.4 follow up visits) applying appropriate national average hourly wage rates . than were specialist patients (mean of 2.8 follow-up visits) (differ- The cost questionnaire formed part of the questionnaire package ence 0.6; 95% CI 0.3Ð0.9; P < 0.001). Each follow-up visit was containing the instruments measuring HRQOL, described previ- longer in the general practice group (mean 10.5 min) than in the ously (Grunfeld et al, 1996). The questionnaire package was hospital group (mean 7.4 min) (difference 3.1; 95% CI 2.6Ð3.6; posted to study participants at three points during the 18-month P < 0.001), based on physiciansÕ reports on the record-of-visit study period: baseline (before randomization when all follow-up form. Similarly, over the 18 months, the total time for follow-up appointments were in hospital outpatient clinics), midtrial and at visits was longer in the general practice group (mean 35.6 min) the end of the trial. One reminder was sent if the questionnaire had than in the hospital group (mean 20.7 min) (difference 14.9; 95% not been returned within 2 weeks. Participants completed the CI 11.3Ð18.4; P < 0.001). midtrial questionnaire within 10 days of a follow-up visit. As there were no important differences in the results obtained at midtrial Health service costs per patient from those obtained at the end of the trial, midtrial results are The mean total cost and cost per visit was significantly less per reported here because they were more closely linked to a recent patient in the general practice group than in the hospital group follow-up appointment. (Table 1). Although there was no difference between groups in the overall cost of diagnostic tests per patient (Table 1), there were *An assurance of confidentiality was requested and given to the study hospitals significant differences between groups in the costs of individual regarding unit costs. For this reason, only the average of the costs for the study diagnostic tests (Table 2). General practitioners ordered more of hospitals is given. particular diagnostic tests than did specialists, resulting in signifi-  National average wage rates for manual and non-manual labour taken from New Earnings Survey (1994). cantly greater costs for chest radiographs, blood tests [full blood British Journal of Cancer (1999) 79(7/8), 1227–1233 © Cancer Research Campaign 1999 Economic evaluation of breast cancer follow-up 1229 Table 1 Average costs (£s) per patient by trial group GP group Hospital group Difference n = 148 n = 141 GP – Hospital Resource item Mean (s.d.) Range Mean (s.d.) Range (95% CI) Cost of visits 40.9 5.8–143.8 174.1 62.0–558.0 –133.2* (20.1) (85.1) (–147.8;–118.7) Cost of tests 23.8 0.0–158.2 20.9 0.0–204.9 2.9 (29.7) (36.3) (–4.8; 10.6) Total costs 64.7 5.8–301.9 195.1 62.0–737.4 –130.4* (42.8) (107.4) (–149.1;–111.6) *P < 0.001. Table 2 Mean frequency and cost of each type of test by trial group Test Unit Mean frequency and costs (£s) cost GP group Hospital group Difference (£s) No./Pt Cost/Pt No./Pt Cost/Pt (95% CI) P-Value –0.001 0.959 Biopsy 19.52 0.02 0.39 0.021 0.42 (–0.039; 0.037) –0.045 0.223 Bonescan 79.08 0.054 4.27 0.099 7.83 (–0.118; 0.028) 0.105 0.007 Chest radiograph 13.17 0.162 2.13 0.057 0.75 (0.029; 0.182) 0.112 0.002 Full blood count 5.47 0.148 0.81 0.036 0.19 (0.042; 0.182) 0.087 0.002 Liver enzymes 6.84 0.108 0.74 0.021 0.14 (0.031; 0.143) 0.138 0.043 Mammograms 26.37 0.493 13.00 0.355 9.36 (0.004; 0.273) –0.029 0.188 Needle biopsy (FNA) 21.76 0.014 0.30 0.043 0.93 (–0.072; 0.014) 0.007 0.592 Abdominal ultrasound 23.15 0.014 0.32 0.007 0.16 (–0.017; 0.03) 0.011 0.720 Radiograph (other) 20.32 0.068 1.38 0.057 1.15 (–0.049; 0.07) 0.054 0.004 ESR 9.00 0.054 0.49 0 0 (0.017; 0.091) a b GP group, n = 148; hospital group, n = 141; unit costs are the average of the unit costs provided by the study hospitals for each diagnostic test. count (FBC), erythrocyte sedimentation rate (ESR) and liver no more than four patients in either group incurred lost income, and enzymes] and mammograms (Table 2). only three patients in the hospital group incurred costs for child care (Table 4). Significantly more patients in the general practice group Health service costs per visit were able to walk to their appointment (Table 4), and significantly Over the 18 months of the study, there was a total of 501 follow-up more patients in the hospital group incurred out-of-pocket expenses visits in the general practice group and 397 visits in the hospital for car parking (Table 4). With respect to accompanying persons, group. The general practice patients were seen significantly more there were no differences between general practice and hospital often than were the specialist patients. Although the mean cost per group in the proportion who took time off work [4 out of 140 (2.9%) visit in general practice was significantly less, the mean cost of vs 9 out of 126 (7.1%) (P = 0.21) respectively] or lost wages [1 out diagnostic tests per visit was similar in the two groups (Table 3). of 140 (0.7%) vs 3 out of 126 (2.4%) (P = 0.41) respectively]. The lower cost in general practice was attributable to lower physi- Patients in the general practice group spent significantly less ciansÕ costs per visit (Table 3). time getting to and from their appointment and waiting to see the doctor than patients in the hospital group. Patients in the general practice group, however, spent significantly more time with the Patient costs doctor than patients in the hospital group (Table 5). For example, There were more patients in paid employment in the general practice based on patientsÕ reports in the cost questionnaire, an average group than in the hospital group (Table 4). For those patients in paid follow-up appointment in the general practice group took a total of employment, more in the hospital group took time off work, 52.6 min (of which 12.8 min were spent with the doctor) compared although there was no difference between groups in the proportion with 82.2 min (of which 6.9 min were spent with the doctor) in the losing wages to attend the follow-up appointment (Table 4). Of note, hospital group (Table 5). © Cancer Research Campaign 1999 British Journal of Cancer (1999) 79(7/8), 1227–1233 1230 E Grunfeld et al Table 3 Average cost (£s) per visit by trial group GP group Hospital group Difference n = 501 n = 397 GP – hospital Resource item Mean (s.d.) Range Mean (s.d.) Range (95% CI) Cost of physician 12.1 3.5–51.8 62.0 62.0–62.0 –49.9* (4.6) (–50.4;–49.5) Cost of tests 6.9 0.0–105.5 7.2 0.0–111.8 –0.3 (14.8) (17.8) (–2.4; 1.9) Total cost per visit 19.0 3.5–128.5 69.2 62.0–173.8 –50.2* (16.6) (17.8) (–52.5; –47.9) *P < 0.001. Table 4 Midtrial patient resource events by trial group GP group Hospital group c c n = 140 n = 126 Parameter No. % No. % P-Value Employed 65 47.8 36 31.0 0.023 Not employed 71 52.2 80 69.0 Took time off work 21 32.3 22 61.1 0.006 No time off work 44 67.7 14 38.9 Lost wages 3 4.6 4 11.4 0.24 No lost wages 62 95.4 31 88.6 Transportation to appointment Walk 45 32.4 2 1.6 Bus 11 7.9 18 14.8 0.000 Car 80 57.6 97 79.5 Other 3 2.2 5 4.1 Out-of-pocket expenses Yes 3 2.4 12 11.0 0.008 No 121 97.6 97 89.0 Need for child care Yes – – 3 2.6 0.06 No 135 100.0 114 97.4 As the same mode of transportation was used for journey to and from appointment, only data for transportation to appointment b c shown. In all cases, out-of-pocket expenses were for car parking. For some parameters No. does not sum to ‘n’ because of missing data. Table 5 Midtrial patient time by trial group Parameter Group No. Time (min) Difference in means Mean (s.d.) (95% CI) Time to get to appointment GP 135 13.1 (8.3) –13.56* Hospital 120 26.7 (15.9) (–16.65; –10.37) Time to get back from appointment GP 138 13.6 (8.6) –14.12* Hospital 118 27.7 (15.3) (–17.12; –11.12) Time waiting to see the doctor GP 136 13.0 (10.7) –10.3* Hospital 119 23.3 (19.9) (–14.2; –6.5) Time with the doctor GP 138 12.8 (5.8) 5.9* Hospital 121 6.91 (4.1) (4.7; 7.2) GP 131 52.6 (22.1) –29.6* Total time for appointment Hospital 115 82.2 (31.8) (–36.5; –22.8) *P < 0.001. British Journal of Cancer (1999) 79(7/8), 1227–1233 © Cancer Research Campaign 1999 Economic evaluation of breast cancer follow-up 1231 Table 6 Average costs for 18 months of follow-up by trial group GP group costs (£s) Hospital group (£s) n = 148 n = 148 (average number of visits = 3.39) (average number of visits = 2.81) Direct medical costs Diagnostic tests 3526.84 3100.60 Physician consultations 6053.20 25771.24 Direct non-medical costs Patient travel 817.80 748.58 Out-of-pocket expenses 7.12 301.12 Time costs Lost wages Patient 62.29 107.18 Accompanying person 20.76 80.38 Total 10488.01 30109.10 a b Patient costs based on data from midtrial assessment. Calculation based on reported travel costs to clinics and GP offices uprated to 1994 prices. Average cost to GP office £1.63; average cost to clinic £1.80, assuming approximate proportion of patients driving to GP office is 42% and driving to clinic is 77% (Wilson, 1991). Travel cost per visit multiplied by number of patients and average number of visits per patient in each group Calculation based on out-of-pocket expenses reported by patients at midtrial multiplied by average number of visits per patient in each group. Valuation of leisure time and work time costs (other than lost wages) not included in this analysis. Calculation based on average total time costs of patients at midtrial (gross hourly wage for non-manual labour ´ average total time for appointment) multiplied by number of patients reporting lost wages (for self or accompanying person) multiplied by average number of visits per patient in each group. Wages based on 1994 gross wages for non-manual employment (almost all patients reporting lost wages were in non-manual labour). (New Earnings Survey, 1994.) Cost minimization analysis diagnostic testing during the routine follow-up period. Hence, this study provides purchasers in district health authorities with the The total direct health care and patient costs and time costs for the means to calculate for themselves the costs associated with the study cohort during the 18 months of follow-up are reported in different follow-up strategies. Although a formal sensitivity Table 6. Despite fewer visits per patient in the hospital group, the analysis was not conducted, even reducing the unit cost of an large difference in direct medical costs between groups was due to outpatient consultation by 50% would have minimal impact on the the greater cost of specialist outpatient visits. The difference in overall findings of the cost analysis. direct non-medical costs to patients was due to higher costs The cost calculations are based on average costs rather than incurred by patients in the hospital group for car parking. Patient marginal costs because average costs provide a better estimate of travel costs (other than car parking) were similar between groups the cost of services affecting a large number of facilities because the larger costs incurred by patients in the hospital group (Drummond and Jefferson, 1996). Thus, the use of average costs were offset by fewer follow-up visits. Time costs in the form of enhances the generalizability of the results (Drummond and wages foregone for patients and accompanying persons were Jefferson, 1996), but may overestimate the costs in the short-term greater in the hospital group because of the greater time taken to when compared with a calculation based on marginal costs attend a specialist outpatient visit. This was the case despite (Robinson, 1993). similar proportions in each group taking time off work and fewer An important finding of this study is that general practitioners follow-up visits in the hospital group. ordered more diagnostic tests than specialists. GPs were provided with guidelines that mammograms were the only investigation DISCUSSION recommended routinely. Nevertheless, they ordered five times as many FBCs and liver enzymes, three times as many chest radi- Health service costs ographs, and eight times as many ESRs. This finding is not The calculation of health service costs is based on unit costs for surprising because it has been suggested that GPs might order diagnostic tests and outpatient consultations obtained directly diagnostic tests more frequently than specialists because of lack of from the institutions involved in the study, both of which were confidence in their clinical skills. One other possible explanation district general hospitals in central England. Uncertainty exists as is that these tests were ordered because of a co-morbidity for to whether the unit costs provided by the institutions reflect the which the GP was also following the patient. real costs of the services, and whether they are typical of costs No previous study has compared the cost of specialist follow-up elsewhere in England. The variation in costs in different districts in with follow-up in primary care. Other reports have compared England (and elsewhere) may result in less cost savings for some ÔintensiveÕ with ÔminimalistÕ follow-up strategies (Schapira, 1993; institutions and more for others. To facilitate comparison with Mapelli et al, 1995; Virgo et al, 1995). Although the ÔminimalistÕ other settings, data on the quantity as well as the cost of health follow-up strategy in those reports is similar to the control arm of service resource use are provided (Drummond and Davies, 1991; this study (periodic examination and routine mammograms), the Drummond and Jefferson, 1996). This study provides data on experimental arm goes one step further by providing ÔminimalistÕ the average frequency and length of visits, and frequency of follow-up in primary care. The results of the economic evaluation © Cancer Research Campaign 1999 British Journal of Cancer (1999) 79(7/8), 1227–1233 1232 E Grunfeld et al of this study show that the lower cost of the experimental strategy CONCLUSIONS is primarily attributable to the lower cost of a general practitioner The RCT showed no increase in the clinical outcomes of delay in consultation compared with a specialist consultation in hospital diagnosing recurrence and reinitiating specialist care as a result of clinics. This cost saving, however, might be counterbalanced by primary care follow-up (Grunfeld et al, 1996). Process measures the documented increased use of diagnostic tests by GPs, and the of the quality of clinical care such as frequency and length of visits expected (but not documented) further costs of false-positive test (Health Services Research Group, 1992) were superior in primary results. Before it can be concluded with certainty that primary care care. follow-up is the less costly follow-up strategy, further investiga- The cost of physician visits and patient costs were lower in tion will need to be carried out. primary care. However, there was no difference in the total costs of diagnostic tests, with particular tests being performed more Patients costs frequently in primary care than in specialist care. Thus, the lower health service costs of primary care follow-up were attributable to In this study, data were collected on direct costs to patients (travel, the lower cost of a physician visit. The impact of excess costs of child care, out-of-pocket expenses), time costs (wages lost to tests and of false-positive test results would have to be evaluated patients and accompanying persons), and the time taken for a before it can be concluded with certainty that primary care follow- follow-up visit. Lost wages were valued according to national up is the less costly option overall. These excess costs may average wage rates. The time taken for a follow-up visit, however, counterbalance the lower costs of a visit in countries where the was significantly greater in the hospital group than in the GP cost differential between primary care physicians and specialist group. In this study, no monetary value was given for the time physicians is not as great as it is in the UK. taken for a follow-up visit. Although it is recognized that there is an opportunity cost associated with time forgone to attend a follow-up visit (Robinson, 1993; Sculpher and Buxton, 1993), ACKNOWLEDGEMENTS valuation of this time was not undertaken because of the un- We thank Jean Pugh, Jo Horler, Christine Southwell and Sally resolved controversy over the best way to value leisure time costs Black for their excellent and dedicated work as research nurses; and non-working time costs (Drummond et al, 1997). Further Drs John Clements and Graham Cradduck for advice and support research is warranted to model Ð according to the different in general practitioner liaison; and Karen Lee for data analysis. We approaches to valuing time Ð the private time costs associated with thank the many surgeons, oncologists and general practitioners the different follow-up strategies. The high prevalence of breast whose commitment made this research possible. We particularly cancer and the length of current follow-up practices suggests that thank the women with breast cancer who participated in this private time costs will have a significant impact when viewed research. from the societal perspective (Sculpher and Buxton, 1993; EG is supported in part by the Ontario Ministry of Health. At Drummond et al, 1997). The finding of this study that significantly the time of the research, EG was a fellow of the National Cancer more patients walked to general practice visits is consistent with Institute of Canada supported with funds from the Canadian other studies comparing GP and hospital-based screening clinics Cancer Society. The research was funded by the Department of (Sculpher and Buxton, 1993). Similarly, the finding that a minority Health for England and Wales with a generous contribution from of patients in either group took time off work and a very small the Ballakermean School of the Isle of Man and support from the minority lost wages is consistent with other studies (Sculpher and General Practice Research Group of the Imperial Cancer Research Buxton, 1993). Fund. In a comparative study of specialist vs. shared primary care/specialist paediatric cancer care, the greatest proportion of saved costs was associated with patient non-medical costs in the form of travel costs, opportunity costs of time and lost wages REFERENCES (Strayer et al, 1980). When one therapy results in better outcomes Breast Cancer Surveillance Expert Panel (1997) Recommended breast cancer than another, patients will choose the best therapy regardless of surveillance guidelines. J Clin Oncol 15: 2149Ð2156 costs. When clinical outcomes are equivalent, however, personal Clark PB and Morris DL (1981) Management of patients after mastectomy. Br Med costs will influence patientsÕ decisions about health care alter- J 282: 2095Ð2096 Dewar J (1995) Follow-up in breast cancer: a suitable case for reappraisal. Br Med J natives (Strayer et al, 1980). This may be particularly true for 310: 685Ð686 screening programmes in which benefits are less tangible and Donegan WL (1995) Follow-up after treatment for breast cancer: how much is too personal costs may prove to be economic barriers to the use of much? J Surg Oncol 59: 211Ð214 such programmes (Sculpher and Buxton, 1993). The possibility Drummond MF and Davies LM (1991) Economic analysis alongside clinical trials. that a proportion of patients may not be attending regular follow- Int J Technol Assess Health Care 7: 561Ð573 Drummond MF and Jefferson TO (1996) Guidelines for authors and peer reviewers up in specialist clinics because of economic barriers has never of economic submissions to the BMJ. Br Med J 313: 275Ð283 been raised in the literature on follow-up. (As noted previously, Drummond MF, OÕBrien B, Stoddart GL and Torrance GW (1997) Methods for the follow-up can be viewed as a screening programme for recurrent Economic Evaluation of Health Care Programmes, 2nd edn. Oxford University disease.) Although the data from this study cannot illuminate this Press: Oxford Garner SB, Winter PD and Garner MJ (1991) CIA statistical program version 1.1 point (as study patients were already regularly attending for GIVIO Investigators (1994) Impact of follow-up testing on survival and health- follow-up), it is an interesting possibility that GP follow-up related quality of life in breast cancer patients. JAMA 271: 1587Ð1592 may make the follow-up programme available to a broader Grunfeld E, Mant D, Yudkin P, Adewuyi-Dalton R, Cole D, Stewart J, Fitzpatrick R range of patients, who may otherwise be deterred by economic and Vessey MPV (1996) Routine follow-up of breast cancer in primary care: a barriers. randomised trial. Br Med J 313: 665Ð669 British Journal of Cancer (1999) 79(7/8), 1227–1233 © Cancer Research Campaign 1999 Economic evaluation of breast cancer follow-up 1233 Health Services Research Group (1992) Quality of care. 1. What is quality and how Sculpher MJ and Buxton MJ (1993) The Private Costs Incurred when Patients can it be measured? Can Med Assoc J 146: 2153Ð2158 Visit Screening Clinics: The Cases of Screening for Breast Cancer and for Hughes LE and Courtney SP (1985) Follow-up of patients with breast cancer. Br Diabetic Retinopathy, pp.1Ð18. Health Economics Group: Brunel Med J 290: 1229Ð1230 University Loprinzi C (1995) Follow-up testing for curatively treated cancer survivors. JAMA Strayer F, Kisker CT and Fethke C (1980) Cost-effectiveness of a shared 273: 1877Ð1878 management delivery system for the care of children with cancer. Pediatrics Mapelli V, Dirindin N and Grilli R (1995) Economic evaluation of diagnostic follow- 66: 907Ð911 up after primary treatment for breast cancer. Ann Oncol 6: S61ÐS64 Tomin R and Donegan WL (1987) Screening for recurrent breast cancer Ð its McWhinney IR, Hoddinott SN, Bass MJ, Gay K and Shearer R (1990) Role of the effectiveness and prognostic value. J Clin Oncol 5: 62Ð67 family physician in the care of cancer patients. Can Fam Phys 36: 2183Ð2186 Virgo KS, Vernava AM, Longo WE, McKirgan LW and Johnson FE (1995) Cost of Robinson R (1993) Costs and cost-minimisation analysis. Br Med J 307: 726Ð728 patient follow-up after potentially curative colorectal cancer treatment. JAMA Rosselli Del Turco M, Palli D, Cariddi A, Ciatto S, Pacini P and Distante V (1994) 273: 1837Ð1841 Intensive diagnostic follow-up after treatment of primary breast cancer. JAMA Wilson A (1991) Consultation length in general practice: a review. Br J Gen Pract 271: 1593Ð1597 41: 119Ð122 Schapira DV (1993) Breast cancer surveillance Ð a cost-effective strategy. Breast Worster A, Wood ML, McWhinney IR and Bass MJ (1995) Who provides follow-up Cancer Res Treat 25: 107Ð111 care for patients with breast cancer? Can Fam Phys 41: 1314Ð1319 © Cancer Research Campaign 1999 British Journal of Cancer (1999) 79(7/8), 1227–1233 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png British Journal of Cancer Springer Journals

Follow-up of breast cancer in primary care vs specialist care: results of an economic evaluation

Loading next page...
 
/lp/springer-journals/follow-up-of-breast-cancer-in-primary-care-vs-specialist-care-results-fe0qoh940b

References (33)

Publisher
Springer Journals
Copyright
Copyright © 1999 by The Author(s)
Subject
Biomedicine; Biomedicine, general; Cancer Research; Epidemiology; Molecular Medicine; Oncology; Drug Resistance
ISSN
0007-0920
eISSN
1532-1827
DOI
10.1038/sj.bjc.6690197
Publisher site
See Article on Publisher Site

Abstract

British Journal of Cancer (1999) 79(7/8), 1227–1233 © 1999 Cancer Research Campaign Article no. bjoc.1998.0197 Follow-up of breast cancer in primary care vs specialist care: results of an economic evaluation 1,2,6 4 1,5 1 1 6 7 3 1 E Grunfeld , A Gray , D Mant , P Yudkin , R Adewuyi-Dalton , D Coyle , D Cole , J Stewart , R Fitzpatrick and M Vessey 1 2 Division of Public Health and Primary Health Care, Institute of Health Sciences, University of Oxford, Oxford OX3 7LF, UK; Ottawa Regional Cancer Centre, Department of Medicine, University of Ottawa, Ontario K1H 8L6, Canada; Northamptonshire Centre for Oncology, Northampton General Hospital, Northampton 4 5 NN1 5BD, UK; Health Economics Research Centre, Institute of Health Sciences, University of Oxford, Oxford OX3 7LF, UK; Department of Primary Medical Care, University of Southampton, Southampton SO1 6ST, UK; Clinical Epidemiology Unit, Loeb Research Institute, University of Ottawa, Ontario K1Y 4E9, Canada; Princess Margaret Hospital, Swindon SN1 4JU, UK Summary A randomized controlled trial (RCT) comparing primary-care-centred follow-up of breast cancer patients with the current standard practice of specialist-centred follow-up showed no increase in delay in diagnosing recurrence, and no increase in anxiety or deterioration in health-related quality of life. An economic evaluation of the two schemes of follow-up was conducted concurrent with the RCT. Because the RCT found no difference in the primary clinical outcomes, a cost minimization analysis was conducted. Process measures of the quality of care such as frequency and length of visits were superior in primary care. Costs to patients and to the health service were lower in primary care. There was no difference in total costs of diagnostic tests, with particular tests being performed more frequently in primary care than in specialist care. Data are provided on the average frequency and length of visits, and frequency of diagnostic testing for breast cancer patients during the follow-up period. Keywords: breast cancer; economic evaluation; follow-up care; randomized controlled trial; primary care After completing primary treatment, it is standard practice in most 1996). We report here the results of the economic evaluation to countries for breast cancer patients to be followed in specialist assess the relative costs of the two alternative schemes of follow- outpatient clinics during the disease-free interval. This practice of up that was conducted concurrent with the RCT. Because the RCT post-treatment surveillance has recently come under close scrutiny showed no important differences in the primary clinical outcomes, (Dewar, 1995; Donegan, 1995; Loprinzi, 1995; Breast Cancer the form of economic evaluation was cost minimization whereby Surveillance Expert Panel, 1997). Studies to evaluate its effective- the two schemes are examined for the least costly alternative ness have used a Ôless intensiveÕ specialist follow-up regimen as (Drummond et al, 1997). the comparator (GIVIO Investigators, 1994; Rosselli Del Turco et al, 1994). No previous studies have evaluated prospectively the cost effectiveness of this practice of long-term follow-up of breast PARTICIPANTS AND METHODS cancer patients. Participants and methods for the RCT We conducted a randomized controlled trial (RCT) comparing specialist follow-up with follow-up by the patientÕs own general Participants were 296 women with breast cancer in remission practitioner (Grunfeld et al, 1996). The results showed no increase receiving regular follow-up at two district general hospitals in in delay in diagnosing recurrence and reinitiating specialist care as England. These women were taking part in an RCT to evaluate a a result of primary care follow-up. As has been reported in primary-care-based system of routine breast cancer follow-up, previous studies (Clark and Morris, 1981; Hughes, 1985; Tomin whereby they were randomized to one of two groups: continued and Donegan, 1987; McWhinney et al, 1990; Worster et al, 1995), routine follow-up in hospital outpatient clinics according to usual the results of the RCT (Grunfeld et al, 1996) showed that most practice (hospital group) or routine follow-up from their own recurrences are detected in the interval between regularly sched- general practitioner (general practice group). Ethical approval to uled follow-up appointments, and many are presented to the conduct the study was obtained from the local research ethics general practitioner irrespective of the formal follow-up arrange- committees. The recommended follow-up regimen was the same ments. The results also showed no increase in anxiety or deteriora- for women in both groups and involved periodic physician visits tion in health-related quality of life (HRQOL) (Grunfeld et al, for a breast cancer check-up, routine surveillance mammograms (the frequency depended on initial treatment and age), and diag- Received 18 May 1998 nostic testing only if clinically indicated. A full description of Accepted 30 July 1998 study participants, study methods and results of the primary outcomes of time to diagnosis of recurrence and HRQOL have Correspondence to: E Grunfeld, Ottawa Regional Cancer Centre, been reported previously (Grunfeld et al, 1996). Data collection Department of Medicine, University of Ottawa, 501 Symth Road, Ottawa, Ontario K1H 8L6, Canada for economic evaluation was an integral part of the RCT. 1227 1228 E Grunfeld et al Measurement and analysis of costs Analysis The two-tailed t-test was used to assess the significance of The perspective of the economic evaluation considered costs to the between group differences in means. The c test was used to assess health service (particularly, the costs of follow-up visits and diag- the significance of between group differences in proportions. nostic tests) and costs to the patients (such as lost earnings and out- Statistical analysis was performed using the software package of-pocket expenses). Valuation of resources was based on unit Statistical Package for the Social Sciences (SPSS, version 6.1.2). costs from the providers and national averages. All costs are Confidence intervals were calculated using the statistical program expressed in 1994 UK £s. Discounting was not considered to be CIA (Garner, 1991). All analysis was on an intention-to-treat relevant to this analysis, which concerns an 18-month time period. basis. Health service costs Information on health service resource use was collected prospec- Response rates tively at both hospital and general practice visits by means of a Completed record-of-visit forms were returned by physicians for record-of-visit form. This was completed at all consultations 100% (148 out of 148) of patients in the general practice group possibly related to breast cancer by doctors for patients in both (i.e. at least one form was received for each patient) and for 95.3% groups and provided details of the frequency of visits, the duration (141 out of 148) of patients in the hospital group. of each visit, and the diagnostic tests ordered at each visit. The The baseline cost questionnaire was completed by 99.3% (147 study hospitals provided unit costs for each type of test and for an out of 148) of patients in the general practice group and 95.3% average outpatient consultation (based on the total annual running (141 out of 148) of patients in the hospital group. By midtrial, four and capital costs of the department averaged across the total patients had died in the general practice group and five had died in number of attendances); the equivalent total costs for an average the hospital group. The adjusted response rate for the midtrial cost general practice consultation was taken from national averages. questionnaire was 97.2% (140 out of 144) and 88.7% (126 out of In the analysis of costs, the average of the unit costs for the 142) respectively. At the end of the study period, a total of six diagnostic tests, procedures and outpatient consultations provided patients had died in the general practice group and one had moved by the two study hospitals was used.* For estimates of the general and so was lost to follow-up, whereas a total of 12 patients had practitioner consultations, UK Department of Health estimates of died in the hospital group and one had moved and so was lost to the cost of a consultation were adjusted to allow for the differences follow-up. The adjusted response rate for the end of the trial cost in the average length of surgery and home consultations (Wilson, questionnaire was 97.2% (137 out of 141) in the general practice 1991). Using this approach, a cost per minute of £1.15 was esti- group and 88.1% (119 out of 135) in the hospital group. mated. To calculate a cost per GP consultation, the cost per minute As the adjusted response rate in the hospital group had fallen to was then applied to the length of each consultation. just above 88% while remaining above 97% in the general practice group, non-respondents in the hospital group were compared on Patient costs baseline characteristics and selected domains of the HRQOL A questionnaire to collect data on costs incurred by the patient was instruments. There were no significant differences between non- developed specifically for this study. The cost questionnaire respondents and respondents on any of these variables. related to costs associated with the patientÕs most recent breast cancer follow-up visit, and included questions about out-of-pocket expenses, lost earnings as a result of time off work, time spent in RESULTS attending the follow-up visit, and costs incurred by accompanying Health service resource use and costs persons. When patients, or those accompanying them, indicated that they had lost earnings over a specified period as a result of a During the 18 months of the study, general practice patients were follow-up visit, the cost to them of this visit was estimated by seen significantly more frequently (mean of 3.4 follow up visits) applying appropriate national average hourly wage rates . than were specialist patients (mean of 2.8 follow-up visits) (differ- The cost questionnaire formed part of the questionnaire package ence 0.6; 95% CI 0.3Ð0.9; P < 0.001). Each follow-up visit was containing the instruments measuring HRQOL, described previ- longer in the general practice group (mean 10.5 min) than in the ously (Grunfeld et al, 1996). The questionnaire package was hospital group (mean 7.4 min) (difference 3.1; 95% CI 2.6Ð3.6; posted to study participants at three points during the 18-month P < 0.001), based on physiciansÕ reports on the record-of-visit study period: baseline (before randomization when all follow-up form. Similarly, over the 18 months, the total time for follow-up appointments were in hospital outpatient clinics), midtrial and at visits was longer in the general practice group (mean 35.6 min) the end of the trial. One reminder was sent if the questionnaire had than in the hospital group (mean 20.7 min) (difference 14.9; 95% not been returned within 2 weeks. Participants completed the CI 11.3Ð18.4; P < 0.001). midtrial questionnaire within 10 days of a follow-up visit. As there were no important differences in the results obtained at midtrial Health service costs per patient from those obtained at the end of the trial, midtrial results are The mean total cost and cost per visit was significantly less per reported here because they were more closely linked to a recent patient in the general practice group than in the hospital group follow-up appointment. (Table 1). Although there was no difference between groups in the overall cost of diagnostic tests per patient (Table 1), there were *An assurance of confidentiality was requested and given to the study hospitals significant differences between groups in the costs of individual regarding unit costs. For this reason, only the average of the costs for the study diagnostic tests (Table 2). General practitioners ordered more of hospitals is given. particular diagnostic tests than did specialists, resulting in signifi-  National average wage rates for manual and non-manual labour taken from New Earnings Survey (1994). cantly greater costs for chest radiographs, blood tests [full blood British Journal of Cancer (1999) 79(7/8), 1227–1233 © Cancer Research Campaign 1999 Economic evaluation of breast cancer follow-up 1229 Table 1 Average costs (£s) per patient by trial group GP group Hospital group Difference n = 148 n = 141 GP – Hospital Resource item Mean (s.d.) Range Mean (s.d.) Range (95% CI) Cost of visits 40.9 5.8–143.8 174.1 62.0–558.0 –133.2* (20.1) (85.1) (–147.8;–118.7) Cost of tests 23.8 0.0–158.2 20.9 0.0–204.9 2.9 (29.7) (36.3) (–4.8; 10.6) Total costs 64.7 5.8–301.9 195.1 62.0–737.4 –130.4* (42.8) (107.4) (–149.1;–111.6) *P < 0.001. Table 2 Mean frequency and cost of each type of test by trial group Test Unit Mean frequency and costs (£s) cost GP group Hospital group Difference (£s) No./Pt Cost/Pt No./Pt Cost/Pt (95% CI) P-Value –0.001 0.959 Biopsy 19.52 0.02 0.39 0.021 0.42 (–0.039; 0.037) –0.045 0.223 Bonescan 79.08 0.054 4.27 0.099 7.83 (–0.118; 0.028) 0.105 0.007 Chest radiograph 13.17 0.162 2.13 0.057 0.75 (0.029; 0.182) 0.112 0.002 Full blood count 5.47 0.148 0.81 0.036 0.19 (0.042; 0.182) 0.087 0.002 Liver enzymes 6.84 0.108 0.74 0.021 0.14 (0.031; 0.143) 0.138 0.043 Mammograms 26.37 0.493 13.00 0.355 9.36 (0.004; 0.273) –0.029 0.188 Needle biopsy (FNA) 21.76 0.014 0.30 0.043 0.93 (–0.072; 0.014) 0.007 0.592 Abdominal ultrasound 23.15 0.014 0.32 0.007 0.16 (–0.017; 0.03) 0.011 0.720 Radiograph (other) 20.32 0.068 1.38 0.057 1.15 (–0.049; 0.07) 0.054 0.004 ESR 9.00 0.054 0.49 0 0 (0.017; 0.091) a b GP group, n = 148; hospital group, n = 141; unit costs are the average of the unit costs provided by the study hospitals for each diagnostic test. count (FBC), erythrocyte sedimentation rate (ESR) and liver no more than four patients in either group incurred lost income, and enzymes] and mammograms (Table 2). only three patients in the hospital group incurred costs for child care (Table 4). Significantly more patients in the general practice group Health service costs per visit were able to walk to their appointment (Table 4), and significantly Over the 18 months of the study, there was a total of 501 follow-up more patients in the hospital group incurred out-of-pocket expenses visits in the general practice group and 397 visits in the hospital for car parking (Table 4). With respect to accompanying persons, group. The general practice patients were seen significantly more there were no differences between general practice and hospital often than were the specialist patients. Although the mean cost per group in the proportion who took time off work [4 out of 140 (2.9%) visit in general practice was significantly less, the mean cost of vs 9 out of 126 (7.1%) (P = 0.21) respectively] or lost wages [1 out diagnostic tests per visit was similar in the two groups (Table 3). of 140 (0.7%) vs 3 out of 126 (2.4%) (P = 0.41) respectively]. The lower cost in general practice was attributable to lower physi- Patients in the general practice group spent significantly less ciansÕ costs per visit (Table 3). time getting to and from their appointment and waiting to see the doctor than patients in the hospital group. Patients in the general practice group, however, spent significantly more time with the Patient costs doctor than patients in the hospital group (Table 5). For example, There were more patients in paid employment in the general practice based on patientsÕ reports in the cost questionnaire, an average group than in the hospital group (Table 4). For those patients in paid follow-up appointment in the general practice group took a total of employment, more in the hospital group took time off work, 52.6 min (of which 12.8 min were spent with the doctor) compared although there was no difference between groups in the proportion with 82.2 min (of which 6.9 min were spent with the doctor) in the losing wages to attend the follow-up appointment (Table 4). Of note, hospital group (Table 5). © Cancer Research Campaign 1999 British Journal of Cancer (1999) 79(7/8), 1227–1233 1230 E Grunfeld et al Table 3 Average cost (£s) per visit by trial group GP group Hospital group Difference n = 501 n = 397 GP – hospital Resource item Mean (s.d.) Range Mean (s.d.) Range (95% CI) Cost of physician 12.1 3.5–51.8 62.0 62.0–62.0 –49.9* (4.6) (–50.4;–49.5) Cost of tests 6.9 0.0–105.5 7.2 0.0–111.8 –0.3 (14.8) (17.8) (–2.4; 1.9) Total cost per visit 19.0 3.5–128.5 69.2 62.0–173.8 –50.2* (16.6) (17.8) (–52.5; –47.9) *P < 0.001. Table 4 Midtrial patient resource events by trial group GP group Hospital group c c n = 140 n = 126 Parameter No. % No. % P-Value Employed 65 47.8 36 31.0 0.023 Not employed 71 52.2 80 69.0 Took time off work 21 32.3 22 61.1 0.006 No time off work 44 67.7 14 38.9 Lost wages 3 4.6 4 11.4 0.24 No lost wages 62 95.4 31 88.6 Transportation to appointment Walk 45 32.4 2 1.6 Bus 11 7.9 18 14.8 0.000 Car 80 57.6 97 79.5 Other 3 2.2 5 4.1 Out-of-pocket expenses Yes 3 2.4 12 11.0 0.008 No 121 97.6 97 89.0 Need for child care Yes – – 3 2.6 0.06 No 135 100.0 114 97.4 As the same mode of transportation was used for journey to and from appointment, only data for transportation to appointment b c shown. In all cases, out-of-pocket expenses were for car parking. For some parameters No. does not sum to ‘n’ because of missing data. Table 5 Midtrial patient time by trial group Parameter Group No. Time (min) Difference in means Mean (s.d.) (95% CI) Time to get to appointment GP 135 13.1 (8.3) –13.56* Hospital 120 26.7 (15.9) (–16.65; –10.37) Time to get back from appointment GP 138 13.6 (8.6) –14.12* Hospital 118 27.7 (15.3) (–17.12; –11.12) Time waiting to see the doctor GP 136 13.0 (10.7) –10.3* Hospital 119 23.3 (19.9) (–14.2; –6.5) Time with the doctor GP 138 12.8 (5.8) 5.9* Hospital 121 6.91 (4.1) (4.7; 7.2) GP 131 52.6 (22.1) –29.6* Total time for appointment Hospital 115 82.2 (31.8) (–36.5; –22.8) *P < 0.001. British Journal of Cancer (1999) 79(7/8), 1227–1233 © Cancer Research Campaign 1999 Economic evaluation of breast cancer follow-up 1231 Table 6 Average costs for 18 months of follow-up by trial group GP group costs (£s) Hospital group (£s) n = 148 n = 148 (average number of visits = 3.39) (average number of visits = 2.81) Direct medical costs Diagnostic tests 3526.84 3100.60 Physician consultations 6053.20 25771.24 Direct non-medical costs Patient travel 817.80 748.58 Out-of-pocket expenses 7.12 301.12 Time costs Lost wages Patient 62.29 107.18 Accompanying person 20.76 80.38 Total 10488.01 30109.10 a b Patient costs based on data from midtrial assessment. Calculation based on reported travel costs to clinics and GP offices uprated to 1994 prices. Average cost to GP office £1.63; average cost to clinic £1.80, assuming approximate proportion of patients driving to GP office is 42% and driving to clinic is 77% (Wilson, 1991). Travel cost per visit multiplied by number of patients and average number of visits per patient in each group Calculation based on out-of-pocket expenses reported by patients at midtrial multiplied by average number of visits per patient in each group. Valuation of leisure time and work time costs (other than lost wages) not included in this analysis. Calculation based on average total time costs of patients at midtrial (gross hourly wage for non-manual labour ´ average total time for appointment) multiplied by number of patients reporting lost wages (for self or accompanying person) multiplied by average number of visits per patient in each group. Wages based on 1994 gross wages for non-manual employment (almost all patients reporting lost wages were in non-manual labour). (New Earnings Survey, 1994.) Cost minimization analysis diagnostic testing during the routine follow-up period. Hence, this study provides purchasers in district health authorities with the The total direct health care and patient costs and time costs for the means to calculate for themselves the costs associated with the study cohort during the 18 months of follow-up are reported in different follow-up strategies. Although a formal sensitivity Table 6. Despite fewer visits per patient in the hospital group, the analysis was not conducted, even reducing the unit cost of an large difference in direct medical costs between groups was due to outpatient consultation by 50% would have minimal impact on the the greater cost of specialist outpatient visits. The difference in overall findings of the cost analysis. direct non-medical costs to patients was due to higher costs The cost calculations are based on average costs rather than incurred by patients in the hospital group for car parking. Patient marginal costs because average costs provide a better estimate of travel costs (other than car parking) were similar between groups the cost of services affecting a large number of facilities because the larger costs incurred by patients in the hospital group (Drummond and Jefferson, 1996). Thus, the use of average costs were offset by fewer follow-up visits. Time costs in the form of enhances the generalizability of the results (Drummond and wages foregone for patients and accompanying persons were Jefferson, 1996), but may overestimate the costs in the short-term greater in the hospital group because of the greater time taken to when compared with a calculation based on marginal costs attend a specialist outpatient visit. This was the case despite (Robinson, 1993). similar proportions in each group taking time off work and fewer An important finding of this study is that general practitioners follow-up visits in the hospital group. ordered more diagnostic tests than specialists. GPs were provided with guidelines that mammograms were the only investigation DISCUSSION recommended routinely. Nevertheless, they ordered five times as many FBCs and liver enzymes, three times as many chest radi- Health service costs ographs, and eight times as many ESRs. This finding is not The calculation of health service costs is based on unit costs for surprising because it has been suggested that GPs might order diagnostic tests and outpatient consultations obtained directly diagnostic tests more frequently than specialists because of lack of from the institutions involved in the study, both of which were confidence in their clinical skills. One other possible explanation district general hospitals in central England. Uncertainty exists as is that these tests were ordered because of a co-morbidity for to whether the unit costs provided by the institutions reflect the which the GP was also following the patient. real costs of the services, and whether they are typical of costs No previous study has compared the cost of specialist follow-up elsewhere in England. The variation in costs in different districts in with follow-up in primary care. Other reports have compared England (and elsewhere) may result in less cost savings for some ÔintensiveÕ with ÔminimalistÕ follow-up strategies (Schapira, 1993; institutions and more for others. To facilitate comparison with Mapelli et al, 1995; Virgo et al, 1995). Although the ÔminimalistÕ other settings, data on the quantity as well as the cost of health follow-up strategy in those reports is similar to the control arm of service resource use are provided (Drummond and Davies, 1991; this study (periodic examination and routine mammograms), the Drummond and Jefferson, 1996). This study provides data on experimental arm goes one step further by providing ÔminimalistÕ the average frequency and length of visits, and frequency of follow-up in primary care. The results of the economic evaluation © Cancer Research Campaign 1999 British Journal of Cancer (1999) 79(7/8), 1227–1233 1232 E Grunfeld et al of this study show that the lower cost of the experimental strategy CONCLUSIONS is primarily attributable to the lower cost of a general practitioner The RCT showed no increase in the clinical outcomes of delay in consultation compared with a specialist consultation in hospital diagnosing recurrence and reinitiating specialist care as a result of clinics. This cost saving, however, might be counterbalanced by primary care follow-up (Grunfeld et al, 1996). Process measures the documented increased use of diagnostic tests by GPs, and the of the quality of clinical care such as frequency and length of visits expected (but not documented) further costs of false-positive test (Health Services Research Group, 1992) were superior in primary results. Before it can be concluded with certainty that primary care care. follow-up is the less costly follow-up strategy, further investiga- The cost of physician visits and patient costs were lower in tion will need to be carried out. primary care. However, there was no difference in the total costs of diagnostic tests, with particular tests being performed more Patients costs frequently in primary care than in specialist care. Thus, the lower health service costs of primary care follow-up were attributable to In this study, data were collected on direct costs to patients (travel, the lower cost of a physician visit. The impact of excess costs of child care, out-of-pocket expenses), time costs (wages lost to tests and of false-positive test results would have to be evaluated patients and accompanying persons), and the time taken for a before it can be concluded with certainty that primary care follow- follow-up visit. Lost wages were valued according to national up is the less costly option overall. These excess costs may average wage rates. The time taken for a follow-up visit, however, counterbalance the lower costs of a visit in countries where the was significantly greater in the hospital group than in the GP cost differential between primary care physicians and specialist group. In this study, no monetary value was given for the time physicians is not as great as it is in the UK. taken for a follow-up visit. Although it is recognized that there is an opportunity cost associated with time forgone to attend a follow-up visit (Robinson, 1993; Sculpher and Buxton, 1993), ACKNOWLEDGEMENTS valuation of this time was not undertaken because of the un- We thank Jean Pugh, Jo Horler, Christine Southwell and Sally resolved controversy over the best way to value leisure time costs Black for their excellent and dedicated work as research nurses; and non-working time costs (Drummond et al, 1997). Further Drs John Clements and Graham Cradduck for advice and support research is warranted to model Ð according to the different in general practitioner liaison; and Karen Lee for data analysis. We approaches to valuing time Ð the private time costs associated with thank the many surgeons, oncologists and general practitioners the different follow-up strategies. The high prevalence of breast whose commitment made this research possible. We particularly cancer and the length of current follow-up practices suggests that thank the women with breast cancer who participated in this private time costs will have a significant impact when viewed research. from the societal perspective (Sculpher and Buxton, 1993; EG is supported in part by the Ontario Ministry of Health. At Drummond et al, 1997). The finding of this study that significantly the time of the research, EG was a fellow of the National Cancer more patients walked to general practice visits is consistent with Institute of Canada supported with funds from the Canadian other studies comparing GP and hospital-based screening clinics Cancer Society. The research was funded by the Department of (Sculpher and Buxton, 1993). Similarly, the finding that a minority Health for England and Wales with a generous contribution from of patients in either group took time off work and a very small the Ballakermean School of the Isle of Man and support from the minority lost wages is consistent with other studies (Sculpher and General Practice Research Group of the Imperial Cancer Research Buxton, 1993). Fund. In a comparative study of specialist vs. shared primary care/specialist paediatric cancer care, the greatest proportion of saved costs was associated with patient non-medical costs in the form of travel costs, opportunity costs of time and lost wages REFERENCES (Strayer et al, 1980). When one therapy results in better outcomes Breast Cancer Surveillance Expert Panel (1997) Recommended breast cancer than another, patients will choose the best therapy regardless of surveillance guidelines. J Clin Oncol 15: 2149Ð2156 costs. When clinical outcomes are equivalent, however, personal Clark PB and Morris DL (1981) Management of patients after mastectomy. Br Med costs will influence patientsÕ decisions about health care alter- J 282: 2095Ð2096 Dewar J (1995) Follow-up in breast cancer: a suitable case for reappraisal. Br Med J natives (Strayer et al, 1980). This may be particularly true for 310: 685Ð686 screening programmes in which benefits are less tangible and Donegan WL (1995) Follow-up after treatment for breast cancer: how much is too personal costs may prove to be economic barriers to the use of much? J Surg Oncol 59: 211Ð214 such programmes (Sculpher and Buxton, 1993). The possibility Drummond MF and Davies LM (1991) Economic analysis alongside clinical trials. that a proportion of patients may not be attending regular follow- Int J Technol Assess Health Care 7: 561Ð573 Drummond MF and Jefferson TO (1996) Guidelines for authors and peer reviewers up in specialist clinics because of economic barriers has never of economic submissions to the BMJ. Br Med J 313: 275Ð283 been raised in the literature on follow-up. (As noted previously, Drummond MF, OÕBrien B, Stoddart GL and Torrance GW (1997) Methods for the follow-up can be viewed as a screening programme for recurrent Economic Evaluation of Health Care Programmes, 2nd edn. Oxford University disease.) Although the data from this study cannot illuminate this Press: Oxford Garner SB, Winter PD and Garner MJ (1991) CIA statistical program version 1.1 point (as study patients were already regularly attending for GIVIO Investigators (1994) Impact of follow-up testing on survival and health- follow-up), it is an interesting possibility that GP follow-up related quality of life in breast cancer patients. JAMA 271: 1587Ð1592 may make the follow-up programme available to a broader Grunfeld E, Mant D, Yudkin P, Adewuyi-Dalton R, Cole D, Stewart J, Fitzpatrick R range of patients, who may otherwise be deterred by economic and Vessey MPV (1996) Routine follow-up of breast cancer in primary care: a barriers. randomised trial. Br Med J 313: 665Ð669 British Journal of Cancer (1999) 79(7/8), 1227–1233 © Cancer Research Campaign 1999 Economic evaluation of breast cancer follow-up 1233 Health Services Research Group (1992) Quality of care. 1. What is quality and how Sculpher MJ and Buxton MJ (1993) The Private Costs Incurred when Patients can it be measured? Can Med Assoc J 146: 2153Ð2158 Visit Screening Clinics: The Cases of Screening for Breast Cancer and for Hughes LE and Courtney SP (1985) Follow-up of patients with breast cancer. Br Diabetic Retinopathy, pp.1Ð18. Health Economics Group: Brunel Med J 290: 1229Ð1230 University Loprinzi C (1995) Follow-up testing for curatively treated cancer survivors. JAMA Strayer F, Kisker CT and Fethke C (1980) Cost-effectiveness of a shared 273: 1877Ð1878 management delivery system for the care of children with cancer. Pediatrics Mapelli V, Dirindin N and Grilli R (1995) Economic evaluation of diagnostic follow- 66: 907Ð911 up after primary treatment for breast cancer. Ann Oncol 6: S61ÐS64 Tomin R and Donegan WL (1987) Screening for recurrent breast cancer Ð its McWhinney IR, Hoddinott SN, Bass MJ, Gay K and Shearer R (1990) Role of the effectiveness and prognostic value. J Clin Oncol 5: 62Ð67 family physician in the care of cancer patients. Can Fam Phys 36: 2183Ð2186 Virgo KS, Vernava AM, Longo WE, McKirgan LW and Johnson FE (1995) Cost of Robinson R (1993) Costs and cost-minimisation analysis. Br Med J 307: 726Ð728 patient follow-up after potentially curative colorectal cancer treatment. JAMA Rosselli Del Turco M, Palli D, Cariddi A, Ciatto S, Pacini P and Distante V (1994) 273: 1837Ð1841 Intensive diagnostic follow-up after treatment of primary breast cancer. JAMA Wilson A (1991) Consultation length in general practice: a review. Br J Gen Pract 271: 1593Ð1597 41: 119Ð122 Schapira DV (1993) Breast cancer surveillance Ð a cost-effective strategy. Breast Worster A, Wood ML, McWhinney IR and Bass MJ (1995) Who provides follow-up Cancer Res Treat 25: 107Ð111 care for patients with breast cancer? Can Fam Phys 41: 1314Ð1319 © Cancer Research Campaign 1999 British Journal of Cancer (1999) 79(7/8), 1227–1233

Journal

British Journal of CancerSpringer Journals

Published: Feb 12, 1999

There are no references for this article.