S.1. the Pan-Canadian Early Detection of Lung Cancer Study

Supplementary Material

S.1. The Pan-Canadian Early Detection of Lung Cancer Study

S.2. The National Lung Cancer Screening Study: Recruitment and Follow-up

S.3. Resource utilization and costs from the Pan-Canadian Early Detection of Lung Cancer study.

Table S.1. Resource utilization and cost analysis

Figure S.1. Source for screening cost and health utility data from the Pan-Canadian Early Detection of Lung Cancer Study

Figure S.2. Source data for the lung cancer treatment cost analysis from the PanCan Early Detection of Lung Cancer Study

S.4. Health utilities

S.5. Societal costs

S.6. PanCan and General risk-NLST transition probabilities

Table S.2. PanCan and General-risk NLST transition probabilities

Figure S.3. Cost-effectiveness acceptability curve

S1. The Pan-Canadian Early Detection of Lung Cancer Study

Costs and quality of life data were prospectively collected from participants in the Pan-Canadian Early Detection of Lung Cancer Study (PanCan; NCT00751660). The PanCan study advertised for recruitment via newspaper, television and radio. Following institutional REB approval at eight major medical centres in Canada, individuals interested in the study were assessed for their risk of developing lung cancer over the telephone. Betweem September, 2008 and December, 2010; potential participants with a calculated risk that was 2% or higher over six years were invited for a baseline LDCT screening test, followed by a second annual screening round for participants without lung nodules. Participants with lung nodules received repeat screening annually for two years. A final, year four screening LDCT was performed on all participants without a diagnosis of lung cancer at that time. Resource utilization and treatment outcomes data were prospectively collected on electronic case report forms until December 31, 2015. The methods for LDCT lung screening in the PanCan study are available online at clinicaltrials.gov, and have been briefly described in the literature.1 Resource utilization data for screening participants (figure S1) and participants receiving treatment for lung cancer (figure S2.) in the PanCan study were used to determine the model cost inputs, according to published methods.2 Preference-based quality of life (health utility) data were ascertained with EQ5D-3L questionnaires administered to PanCan participants before and one year after their initial baseline LDCT screening exam, and used as inputs to the model. Short-term outcomes (lung cancer cases detected, treatment survival and time to curative failure) were also determined to estimate the effect of an alternate screening scenario in the deterministic analysis. At the time of analysis, there was a median duration of follow-up of 5.5 years in the PanCan study.

S2. National Lung Cancer Screening Study: Recruitment and Follow-up

The National Lung Cancer Screening Trial (NLST; NCT00047385) recruited participants between April 2002–April 2004, based on the following selection criteria: 55–74 years of age at the time of randomization and at least 30 pack years of smoking history and having no more than 15 years of ongoing smoking cessation. The NLST data used in this study had a median follow-up time of 6.42 years.

S3. Resource utilization and costs from the Pan-Canadian Early Detection of Lung Cancer study.

The methods of acquiring resource utilization data and assignment of unit costs have been described previously.2 Briefly, resource utilization data relevant to the detection and treatment of lung cancer were collected for all LDCT screening participants and lung cancer patients from eight study centres across Canada between October, 2008 to December 31, 2015. The data were sorted according to their relevance to the detection and treatment of lung cancer or incidental findings and chronologically from the date of the first LDCT baseline screening exam, lung cancer treatment or date of documented disease progression after curative treatment. All costs related to lung cancer detection and treatment were determined from the resource utilization rates reported in table S.2., multiplied by the previously reported unit costs and grouped into specific health states, annually, starting from the baseline screening exam. Participant records were excluded if there was insufficient follow-up to adequately account for resource utilization over the full 365 day cycle in the Markov models for study participants with follow-up, and censored to the last date of resource utilization for participants who were lost to follow-up or deceased. Screening costs were assessed for PanCan participants that were screened with LDCT only (n=1,286). The screening phase costs for PanCan participants screened with LDCT plus autoflourescence bronchoscopy (AFB; n=1,251) were excluded for the purpose of this analysis since AFB screening does not improve early detection and the

costs would be significantly higher than screening undertaken with LDCT alone.2,3

Figure S.1. Source for screening cost and health utility data from the Pan-Canadian Early Detection of Lung Cancer Study

Figure S.2. Source data for the lung cancer treatment cost analysis from the PanCan Early Detection of Lung Cancer Study

Table S.1. Resource utilization and cost analysis for the Pan-Canadian Early Detection of Lung Cancer Study

aPhysician resources include all services for patient consultations and reviews of patient records by medical staff including physicians, dieticians, nurses, social workers and physician group conferences.

S4. Health Utility

Preference based quality of life (health utility) for screening was estimated from EuroQol five-dimension three-level questionnaires (EQ5D-3L) administered to participants in the PanCan study before and after (0-365 days) their baseline screening exam. The average score for 1110 participants who received LDCT screening in the PanCan study before and 365 days after baseline LDCT screening exams were calculated according to preferences of the U.S. population. Quality of life for the treatment and progression health states in the model were estimated from the US-scored EQ5D-3L results that have been published for lung cancer treatment.4,5

S5. Societal costs

Participants at the British Columbia recruiting centre in the PanCan study completed questionnaires about their out-of-pocket expenses to attend screening appointments including travel, childcare, meals and accommodation, net income and lost productivity due to time missed from work to attend screening appointments. A total of 173 participants returned information for at least one screening appointment and the average, per-person costs for each of the three annual screening exams was calculated. Transportation costs were determined based on the average cost per kilometer for standard vehicle use in Canada, parking fees, ferry travel fees at the time of study and self-reported airfare fees. Lost productivity was determined from net income for employed screening participants for the time required to travel to and attend screening appointments. The calculated average cost per visit was used as a per-diem tariff that was applied to each medical visit that the patient was required to attend for screening, once daily ($ per appointment, once per day). First year per-diem tariffs were $76.27 (SE: 24), second year were $87.31 (SE: 15), and third year were $69.00 (SE: 11) for all appointments in the screening health state of the intervention arm. Societal costs for patients receiving lung cancer treatment were obtained by a separate questionnaire following institutional REB approval at the BC Cancer Agency. The questionnaire was administered to consenting lung cancer patients during their oncology appointments (n=88) or after their curative lung cancer treatment (n=10). Lost productivity and caregiver costs were calculated separately for participants receiving non-curative lung cancer treatment. Costs paid by retired and unemployed participants were conservatively excluded from the calculation of annual loss in productivity associated with lung cancer treatment. Caregiver costs were specifically queried based on the amount of unpaid time caregivers took off from work. Treatment phase tariffs were used in both the intervention and comparator arms of the cost-effectiveness analysis. Participants receiving non-curative treatment, including disease progression, had an annual tariff of $15,574 (SE: 2,441), plus a separate per-diem tariff added to all appointments related to cancer treatment (curative or non-curative) equivalent to $59.20 (SE: 19). A calculated annual tariff of $5,532 (SE: 1,907) was added to the first year of curative treatment, one time for participants receiving curative treatments including treatment for curative failures for any screening participant who had surgery for a non-malignant lung nodule. The per-diem tariff for the first year of screening was applied to medical appointments in years 2-30 following curative treatment. The adjusted cost calculations from the societal perspective were used as health state specific cost inputs to the model in the deterministic analysis.

Table S2. PanCan and General-risk NLST transition probabilities

aNon-lung cancer deaths were not verified in the PanCan study

bTransition probabilities for the total NLST with risk criteria available (i.e. High-risk and low-risk subgroups combined)

S6. Incidental findings

The costs of treating incidental findings was estimated from the baseline screening exam for the first 100 PanCan participants receiving LDCT screening at the British Columbia study site. The findings were categorized as actionable or non-actionable and the non-actionable findings were excluded from further analysis. An aggressive, first-line treatment strategy was assigned to each actionable finding and resource utilization rates were multiplied by unit costs. The actionable findings identified in this analysis, and the calculated unit cost are as follows: severe coronary artery calcification ($2,587), breast nodules ($16,281), goiters/enhanced thyroid ($592), severe spinal degenerative changes with progression ($328), tracheal diverticum ($16,790) and ruptured breast implants ($16,000). For each baseline screening exam, there were 0.23 actionable incidental findings reported. The average per-person cost was estimated ($865.56, SE: 337) and added to the first three years of HR-LDCT screening costs, deterministically.

Figure S.2. Cost-effectiveness acceptability curve for the base case scenario of HR-LDCT screening versus standard care.

Supplemental References

1.McWilliams A, Tammemagi M. C., Mayo J. R. et al. Probability of cancer in pulmonary nodules detected on first screening CT. The New England journal of medicine. 2013; 369: 910–9.

2.Cressman S., Lam S., Tammemagi M. C. et al. Resource utilization and costs during the initial years of lung cancer screening with computed tomography in Canada. Journal of thoracic oncology. 2014; 9:1449–58.

3.Tremblay A., Taghizadeh N., McWilliams A. M., et al. Low Prevalence of High Grade Lesions Detected with Autofluorescence Bronchoscopy in the Setting of Lung Cancer Screening in the Pan-Canadian Lung Cancer Screening Study. Chest. 2016; PubMed PMID: 27142184. Epub 2016/05/05.

4.Jang R. W., Isogai P. K., Mittmann N., et al. Derivation of utility values from European Organization for Research and Treatment of Cancer Quality of Life-Core 30 questionnaire values in lung cancer. Journal of thoracic oncology. 2010; 12:1953–7.

5.Tramontano A. C., Schrag D. L., Malin J. K., et al. Catalog and comparison of societal preferences (utilities) for lung cancer health states: results from the Cancer Care Outcomes Research and Surveillance (CanCORS) study. Medical decision making. 2015; 35:371–87.

1