Curve shows the gamma distribution skewing to the right, indicating a small number of people with higher costs. Total mean cost for patients attending the high-risk clinic was A $882 (US $599). The equivalent of US $1 to A $1.47.30
eMethods. Health System Costs
eFigure. High-Risk Clinic Patients Included in the Microcosting Study
eTable. Comparison of Costs for Surveillance Strategies for Individuals at High Risk of Cancer From the Perspective to the Health System
Watts CG, Cust AE, Menzies SW, Coates E, Mann GJ, Morton RL. Specialized Surveillance for Individuals at High Risk for MelanomaA Cost Analysis of a High-Risk Clinic. JAMA Dermatol. 2015;151(2):178-186. doi:10.1001/jamadermatol.2014.1952
Regular surveillance of individuals at high risk for cutaneous melanoma improves early detection and reduces unnecessary excisions; however, a cost analysis of this specialized service has not been undertaken.
To determine the mean cost per patient of surveillance in a high-risk clinic from the health service and societal perspectives.
Design, Setting, and Participants
We used a bottom-up microcosting method to measure resource use in a consecutive sample of 102 patients treated in a high-risk hospital-based clinic in Australia during a 12-month period.
Surveillance and treatment of melanoma.
Main Outcomes and Measures
All surveillance and treatment procedures were identified through direct observation, review of medical records, and interviews with staff and were valued using scheduled fees from the Australian government. Societal costs included transportation and loss of productivity.
The mean number of clinic visits per year was 2.7 (95% CI, 2.5-2.8) for surveillance and 3.8 (95% CI, 3.4-4.1) for patients requiring surgical excisions. The mean annual cost per patient to the health system was A $882 (95% CI, A $783-$982) (US $599 [95% CI, US $532-$665]); the cost discounted across 20 years was A $11 546 (95% CI, A $10 263-$12 829) (US $7839 [95% CI, US $6969-$8710]). The mean annual societal cost per patient (excluding health system costs) was A $972 (95% CI, A $899-$1045) (US $660 [95% CI, US $611-$710]); the cost discounted across 20 years was A $12 721 (95% CI, A $12 554-$14 463) (US $8637 [95% CI, US $8523-$9820]). Diagnosis of melanoma or nonmelanoma skin cancer and frequent excisions for benign lesions in a relatively small number of patients was responsible for positively skewed health system costs.
Conclusions and Relevance
Microcosting techniques provide an accurate cost estimate for the provision of a specialized service. The high societal cost reflects the time that patients are willing to invest to attend the high-risk clinic. This alternative model of care for a high-risk population has relevance for decision making about health policy.
Australia has the highest incidence of melanoma in the world.1 Despite this statistic and the known risk factors for melanoma, screening in the general population is not recommended at present. Selected studies have shown that screening among men older than 50 years as a one-time intervention2 or at 5-year intervals using primary care physicians3 may be cost-effective. The regular surveillance of individuals at high risk for melanoma in a specialized clinic has demonstrated improvements in early detection of lesions4 and a reduction in unnecessary excisions5; however, to our knowledge no studies have examined whether monitoring people at very high risk for developing melanoma for a long period is cost-effective. A specialized high-risk clinic (HRC) for individuals with an elevated risk for melanoma was established within a hospital outpatient clinic at the Sydney Melanoma Diagnostic Centre, Royal Prince Alfred Hospital, Sydney, Australia, in 2006, and this model of care is currently being evaluated and expanded in 3 more centers. Early detection underpins the rationale for surveillance of individuals at high risk for melanoma because the stage at diagnosis affects prognosis6 and the cost of subsequent treatment is lower when melanoma is detected at an early compared with an advanced stage.7,8
The aim of our microcosting study was to measure (1) the cost of providing this specialized skin surveillance service to individuals at very high risk for melanoma and (2) the factors influencing variation in resource use. We aimed to calculate the direct costs of the HRC, that is, costs to the health care system, and the indirect costs not generated directly by the clinic but as a result of attending the clinic. These costs included out-of-pocket costs (eg, travel to the clinic) and the opportunity costs of time forgone in attending the clinic. Microcosting is the reference standard technique9 for calculating the cost of a new service when no published estimates are available.10 Microcosting techniques are preferred for interventions that contain a large component of labor costs and in which interpatient variation in costs is likely, making a mean cost difficult to predict.11 The treatment of patients at high risk for melanoma in a specialized surveillance clinic has been evaluated in only a few countries,12- 15 and, to our knowledge, no studies have reviewed or evaluated the costs of the health care service or the societal costs.
We used a cohort study design. We obtained human research ethics committee approval from the institutional review board of the Royal Prince Alfred Hospital. All participants gave oral consent for one of us (C.G.W.) to attend and observe the HRC consultations. Patients had previously given informed consent for participation in the study.
All patients attending the HRC met at least 1 of the following eligibility criteria: (1) dysplastic nevus syndrome and at least 1 primary invasive melanoma, (2) diagnosis of 2 or more primary invasive melanomas, (3) strong family history of melanoma (≥3 first-degree relatives with melanoma) and 1 or more personal primary invasive melanomas, or (4) presence of the CDKN2A mutation with no requirement for a history of melanoma.16 The strategy used for surveillance in the Sydney HRC has been described previously.12 In brief, this strategy involves regular extended-length consultations once every 6 months, which include a full-body skin examination augmented with dermoscopy and the use of total body photography plus dermoscopy when indicated. When a suspected lesion is identified, the lesion is excised or sequential digital dermoscopic imaging17 of the lesion is commenced and the patient returns for nevus monitoring in 3 months. An additional part of the program is that patients receive instruction in skin self-examination, which they are encouraged to perform using their total body photographs between appointments. A pilot study of this clinic involving 311 patients followed up for a median of 3.5 years showed effective early detection of primary melanoma and a ratio of benign to malignant lesion excision of 1.6:1 for all lesions excised.12
We used a bottom-up microcosting approach to estimate the total costs of HRC care for 12 months. All costs associated with skin surveillance and management of newly identified lesions were included. Procedures related to testing for recurrence of a previously diagnosed cutaneous melanoma, such as chest radiography, were not included because these costs did not apply to HRC surveillance costs. All identifiable direct and indirect costs were included in the analysis and adjusted to 2013 Australian dollars.18 A 20-year time horizon was used to estimate the lifetime costs of surveillance based on the median age of the HRC participants. We applied the Australian standard discount rate of 5%19 to all future costs. The methods followed published guidance on microcosting.10 Additional information is included in the eMethods in the Supplement.
Through direct observation of the HRC from December 10, 2012, through May 14, 2013, we recorded information about each patient’s type and length of consultation (using a stopwatch) and all resource items used. Additional information regarding the mode of transport to the clinic, patient employment status, employment leave type, and requirement of a medical certificate were noted if discussed during the consultation. For employed patients, the occupational group was graded according to the Australian Bureau of Statistics classifications20 to determine the median wage. The presence of accompanying family members or informal caregivers was also recorded.
Staff were interviewed about their roles, the time they spent on various tasks, and their use of consumables. Information about fixed costs and capital and equipment costs was obtained from the HRC records, and prices were checked with equipment suppliers when current receipts were not available. We calculated costs for software licenses, technology maintenance and support, and overhead costs for clinic space. A mean cost per patient for staff in the HRC was derived using salary scales from the Public Health System Awards of New South Wales Health (the state health department).21 Staff costs, including payroll taxes and superannuation, were estimated as an additional 20% to their base salary.22
Sociodemographic data were obtained from the patients’ medical records, and the return travel distance by road from the patient’s home to the clinic was calculated using residential addresses and publicly available software (Whereis; http://www.whereis.com). All documented melanoma surveillance consultations and procedures, whether conducted in the HRC or by other services (eg, primary care), were recorded from the medical records. This information included all diagnostic tests and medication use during the preceding 12 months.
Use of health system resources was calculated for all patients, including surveillance and treatment provided as a result of attending the HRC during the previous 12 months. Resource use items were valued according to the dollar amount subsidized by the Australian government through the Medicare Benefits Schedule.23 The scheduled fees are set annually by the Australian government and provide a value for services against which all residents in Australia can claim a rebate of 100% for primary care services (if the service was billed in bulk) and 85% for non–primary care services. Fees for melanoma surveillance that were reimbursed by the Medicare Benefits Schedule23 were deducted from the hospital salary costs of the dermatology resident to avoid double counting. We allocated a primary care level B (standard-length) or C (extended-length) service, depending on the duration of the consultation. Services provided by primary care physicians in the community were counted as level B consultations. All excised lesions were sent for a pathological examination, with costs varying according to the size and site of the lesion, the number of lesions excised during a consultation, and the complexity of the biopsy material being examined. If a patient was admitted to hospital for an excision, we assigned a health system cost based on the relevant Australian refined diagnosis related groups code.24
A total annual health system cost was calculated for salaries and for overhead and capital costs. A mean cost per consultation was calculated by dividing this total cost by the total number of HRC consultations in 1 year (based on the mean number during the previous 6 years). This figure was used to calculate the mean annual cost per patient.
For patients who reported their mode of travel, we calculated a mean annual travel cost. Based on the reported mode of transport, the number of trips made to the HRC, and the number of surveillance-related consultations documented in the patient’s history, we calculated a total travel cost per patient across 12 months. For air travel, a standard online ticket price was calculated. Hotel accommodation was based on the standard overnight rate for a midrange hotel located within a 5-km radius of the hospital. If individuals had consulted their local primary care physician in the community during the past 12 months for an excision or for removal of sutures related to surveillance at the HRC, a standard return travel distance of 9.6 km (6 miles) was allocated to each primary care visit.25
Documentation in the patient’s medical record was used to calculate the mean out-of-pocket cost for medical treatment related to HRC surveillance in the previous 12 months but not fully rebatable through the Medicare Benefits Schedule23 or the Pharmaceutical Benefits Scheme.26 All patients required an initial set of total body photographs, which were used at each visit for monitoring lesions. At the present time, patients attending the HRC do not pay for these photographs, but this cost likely will be borne by patients in the future. A set of photographs usually lasts 7 years (E.C., personal communication, May 2013); therefore, a mean cost for 7 years was calculated per patient. Mean costs borne by patients for medicines and lotions26 and out-of-pocket costs for specialist services were calculated based on the standard reimbursement through the Medicare Benefits Schedule23 or the Pharmaceutical Benefits Scheme.26
The value of each patient’s time to attend the HRC or to receive other related medical care during the previous 12 months was calculated using the market price of labor (ie, wages or the aged pension).27 Based on our observation, an HRC visit for surveillance or a related procedure required 4 hours or half a day taken from work. For patients who lived more than 100 km (62 miles) away, a full day or 8 hours was required to attend the clinic. For employed patients, we calculated the opportunity cost of time not at work based on an estimated median of full-time weekly total cash earnings by occupation group.28 For patients of working age whose occupation was unknown, we estimated the opportunity cost of lost personal time per visit based on a median of full-time weekly total cash earnings for all Australian employees.28 For patients older than 65 years who were assumed to be retired, a proportion of the weekly single pension was used.29 All wages were adjusted from 2012 to 2013 wage levels by 4.9%, which was the percentage change in full-time mean earnings provided by the Australian Bureau of Statistics.20
All 102 consecutive patients attending the HRC on Mondays from December 10, 2012, through May 14, 2013, were included in the study (eFigure in the Supplement). Of these, 87 were continuing (prevalent) patients and 15 were new (incident) patients attending for the first time. The characteristics of the study population are summarized in Table 1.
Labor in the HRC (ie, staff salaries) was the main component of operational expenses and accounted for 50% of health system costs. Surveillance and procedures accounted for 46% of health system costs; of this group, 235 of 271 consultations (86.7%), excluding specialist consultations, were to the HRC (Table 2). The mean consultation time for a new patient was 40 minutes. All 87 prevalent patients had 2 extended-length consultations during the 12-month period, and these consultations took a mean of 31 (95% CI, 28-34) minutes. Within this group, an additional 66 consultations for nevus monitoring were performed (mean time, 11 [95% CI, 7-15] minutes). Almost half the study group (39 of 87 [45%]) commenced short-term monitoring during the 12-month period, with 1 to 6 lesions identified for short-term monitoring. Seven patients in this group had additional lesions identified for monitoring within the study period. Thirty of the 87 patients (34%) required an excision of a suspected lesion during the 12-month study period; of these, 16 (53%) had 1 lesion excised and 4 (13%) had a range of 4 to 8 excisions. The mean number of HRC visits per year was 2.7 (95% CI, 2.5-2.8) for surveillance and 3.8 (95% CI, 3.4-4.1) for patients requiring surgical excisions.
Three melanomas and 8 keratinocyte carcinomas were detected during the study period (Table 2). Patients had 12 visits to a primary care physician in the community for excision of lesions and 22 visits for removal of excision sutures. Only 1 hospital admission was documented during the study period. Costs per patient were heavily skewed to the right, with a few patients generating costs several times the mean (Figure). The higher health care costs resulted from more excisions of benign lesions with suspected melanoma features and surgical removal of melanoma and nonmelanoma skin cancers.
Mean annual societal costs were very similar to direct health care costs at A $972 (95% CI, A $899-$1045) (US $660 [95% CI, US $611-$710])30 per patient (Table 3). The greatest proportion of these costs was related to the patient’s time spent on surveillance or related activities and travel costs (Table 4). The opportunity cost of time not at work or spent on lost personal time was a mean of A $502 (95% CI, A $425-$579) (US $341 [95% CI, US $290-$388])30 per person during the 12 months. This equates to a cost of A $147 (US $100) each time an individual engages with the health system by attending the HRC or for a related procedure. An opportunity cost for visiting a local primary care physician for removal of sutures was not calculated because we believed the time required would not cause the same disruption to one’s daily schedule. The mean annual out-of-pocket cost for travel was A $117 (US $79) for a consultation or a procedure-related visit, and a wide variation existed in the distance patients traveled to the clinic (Table 4).The annual mean cost for travel and accommodation combined was A $407 (95% CI, A $183-$631 (US $276 [95% CI, US $124-$428])30 per person; the mean cost of traveling by car was 79% of these costs. The mean total number of consultations was less when estimating societal costs because travel time was counted only once when an excision occurred on the same day as a clinic visit. When patients had traveled long distances and an excision was determined at the HRC to be necessary, an attempt was made to organize the surgery on the same day. In our study sample this occurred 6 times.
The out-of-pocket costs for medical treatment were a small component (6%) of patient costs, because only a few visits to specialists outside the HRC were recorded and topical ointments and lotions were infrequently prescribed (Table 4). Visits for specialist procedures related to the location of the excision or patient preference. Two adverse events (wound infections) were noted in the medical records, and both responded to antibiotics. The mean annual and lifetime costs are summarized in Table 3.
The key contributors to the costs of an HRC are labor costs, representing the intensive nature of surveillance, the number of consultations, and the length of time required per patient. The opportunity cost of patients’ time away from work and other activities and the cost of travel are the main drivers for indirect costs.
Recommendations regarding optimal screening (before the diagnosis of melanoma) and surveillance (after the diagnosis) vary within international clinical practice guidelines.31,32 Although population screening is currently not recommended in most countries, some evidence supports follow-up of high-risk groups using an intensive surveillance strategy.33,34 Surveillance of high-risk groups is recommended to continue for longer periods because primary melanomas arise at a higher probability than for the general population during these patients’ lifetimes.34 During the 12-month study period, 34% of the group required an excision of a suspected melanoma lesion and 3 cutaneous melanomas were detected, reflecting the requirement for surveillance.
A systematic review reported significant productivity losses due to the morbidity and premature mortality associated with melanoma.35 Although progress in the treatment of advanced disease has been made, the mean years of life lost owing to metastatic melanoma are greater compared with other cancers (20.4 vs 16.2 years).36 This study quantifies the costs of long-term specialized surveillance in a group with a high probability of future melanoma. The 20-year cost of this model is less than that of other well-established surveillance practices for other cancers37,38 and compares favorably with surveillance costs in the United States for early-stage melanoma8 (eTable in the Supplement). In addition, a substantial cost advantage is gained if melanoma is treated at an early stage. A study of total health care and societal costs across 5 years based on 2008 Medicare reimbursements in the mid-Atlantic states estimated the cost of managing an in situ melanoma at US $5044 compared with the cost of managing a T4b melanoma at US $110 150.8
The model of care assessed in this study uses specialized expertise to maximize the accuracy of melanoma detection at an early stage and to avoid unnecessary excision of benign lesions. However, costs may be lower using other models of care for high-risk individuals, such as primary care practitioners working in general practice or in skin cancer clinics. Barriers for a general practice setting, such as inadequate time to perform full-body skin examinations39 and training in the use of dermoscopy,40 would need to be addressed. Research identifying which groups would benefit most from this specialized screening strategy41,42 will also improve service efficiency.
A small number of patients had more than 4 excisions within 12 months and some had surgery requiring recovery time, which influences health system and out-of-pocket costs. Hospital admissions would affect health system costs; in our study, however, admission was required for only 1% of patients.
A number of limitations of this study have been identified. Only 1 clinic was observed, and other clinics in different locations may vary in their resource use. Although total costs may differ slightly across clinics for patient management and surveillance, the same protocol recommendations for frequency of skin examination, total body photography, and sequential digital dermoscopic imaging apply to all HRCs in the state; the key contributors to costs we identified would also be applicable in other clinics. Similar principles may be applicable internationally. We did not obtain information from all patients; however, our sample captured consecutive patients whom we believe to be typical of patients attending the clinic. The HRC staff attempted to follow up all reports of excisions performed outside the HRC (eg, by primary care physicians in the community), and these excisions were documented in the medical record. However, a small chance exists that some events may not have been captured. Not all pathologic examination reports were cited; where this occurred, we estimated costs based on the summary reports in the patients’ medical records and therefore may have missed costs for additional histologic stains or expert second opinions that would have been detailed in an original report. In this case, we would be more likely to underestimate rather than overestimate the cost of the pathological examination.
We did not calculate the opportunity cost for 4 accompanying adults in this sample because we could not ascertain whether they attended all consultations during the 12-month study period. Finally, we did not calculate lost work days in terms of lost productivity (business cost) because only 1 person in our study had more than 3 consecutive days off work, and this time off was not considered to have a material effect on the overall results. Further research into valuing patient time43 should be considered.
Our study shows that the costs of surveillance for a group of individuals at high risk for melanoma in a specialized HRC can be estimated using microcosting methods. From a health system perspective, the costs of surveillance are driven by the labor costs of the clinic staff and the number of follow-up extended-length surveillance consultations required by these patients. Patients who require more intensive treatment have a greater effect on the overall cost of the program. However, high-risk patients in surveillance programs have been shown to have melanomas detected at an earlier stage compared with high-risk individuals not in a surveillance program.4 Costs for their treatment are therefore likely to have been lower than they would have been in standard care. Treatment at an earlier stage also has advantages in terms of decreased morbidity and cost to the community. Calculating costs using a societal perspective further informs the social cost of surveillance in an HRC. In particular, the costs of opportunity forgone and travel reflect the disease burden and the willingness of patients to travel and give up their own time to attend surveillance, indicating the patient’s perceived acknowledgment of the benefits of attending the HRC. Results from this study will be used in an economic evaluation of the cost-effectiveness of the specialized HRC model of care for individuals at high risk for melanoma compared with standard care in the community. This study will help to inform health policy for melanoma skin screening and follow-up in Australia.
Accepted for Publication: June 23, 2014.
Corresponding Author: Caroline G. Watts, MPH, Cancer Epidemiology and Services Research, Sydney School of Public Health, University of Sydney, 119-143 Missenden Rd, Level 6-North, Camperdown, New South Wales, Australia 2050 (email@example.com).
Published Online: November 12, 2014. doi:10.1001/jamadermatol.2014.1952.
Author Contributions: Ms Watts and Dr Morton had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: All authors.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Watts, Morton.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Watts.
Obtained funding: Cust, Mann.
Administrative, technical, or material support: Cust, Coates, Mann.
Study supervision: Cust, Menzies, Mann, Morton.
Conflict of Interest Disclosures: Dr Cust receives fellowships from the National Health and Medical Research Council and the Cancer Institute NSW. Dr Morton was funded by an Australian National Health and Medical Research Council Sidney Sax–Overseas Public Health Fellowship. No other disclosures were reported.
Funding/Support: This study was supported in part by a doctoral scholarship (Ms Watts), by a fellowship from the Cancer Institute NSW (Dr Cust), and by a Sydney Catalyst Top-Up Research Scholar Award (Ms Watts).
Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: Sharon Lorger, RN, Sydney Medical Diagnostic Center, Sydney Cancer Center, Royal Prince Alfred Hospital, and Ritta Khoury, BMedSci, Sydney Cancer Centre, Royal Prince Alfred Hospital, assisted with high-risk clinic data collection. No compensation was provided for these contributions.