Farquhar et al.[1], England

Clinical data

·    Design: RCT with two arms; randomization by blocks of random size two, four and six, generated by statistician and concealed within sealed opaque envelop until allocation notification by intervention deliverer;

·    Setting: Community setting

·    Sample size: 54 (67 allocated)

·    Recruitment: November 2008-January 2012

·    Data collection: Baseline (t1: week 1 = before randomization), week 3, week 5. Outcomes measured were: patient distress due to breathlessness using a numerical rating scale, disease-specific health related quality-of-life (Chronic Respiratory Questionnaire: CRQ), and anxiety and depression (Hospital Anxiety and Depression Scale: HADS), EQ-5D and measures of service use (8-weeks and 2-weeks prior to baseline and at week3).

Economic evaluation

·    Type: CEA; using primary clinical data

·    Perspective: Not stated; (results for healthcare & social care)

·    Cost year & monetary unit: 2011-2012; GBP

·    Length of evaluation: less than 12-weeks

Funding: Cambridge University Hospitals' NHS Foundation Trust

·    Cancer type: Advanced cancer patients

·    Eligibility criteria: if patients met BIS (Breathlessness Intervention Service) referral criteria (that is, diagnosed appropriately-treated cause of breathlessness, troubled by breathlessness in spite of optimisation of underlying illness, and might benefit from a self-management programme); and not having received BIS previously.

Intervention s: The BIS team comprises: a palliative care medical consultant, a clinical specialist occupational therapist, a clinical specialist physiotherapist and an administrator. At a weekly multidisciplinary team meeting cases are allocated to the most appropriate professional based on information derived from the referral; many patients receive visits from at least two professionals on the team. The intervention is delivered predominantly in the home setting with visits lasting 1-1.5 hours. Visits include interventions relevant to that person and formulation of an individually-tailored exercise plan.

·     IG: intervention group, n=28 (allocated n=35);

·     CWL, control waiting list, n=26 (allocated n=32). Control had to wait and received intervention after week 3.

Program duration : 2-weeks

Variables included in CEA:

·     Patient distress, anxiety, depression and EQ-5D.

·     Healthcare costs, including intervention costs.

·     Informal care costs

Effects:

·     Patient distress due to breathlessness: IG achieved a sign. greater, reduction compared to CWL: adjusted difference −1.29 (95% CI: −2.57 to −0.005), p = 0.049.

·     Incremental QALY-gain of 0.0002 (95%CI: -0.001 to 0.002) for IG vs CWL

·     No sign. differences between arms for other outcomes.

Costs:

·     IG had health/social care savings were on average £211 compared to CWL (95%CI: -£918 to £310).

Economic evaluation :

·     Lower health/social care costs and better primary outcome results for IG indicated dominance over CG

Sensitivity analysis:

·     One-way S.A. were performed. Bootstrapping applied

·     S.A. results confirmed baseline results

Level of evidence (clinical data): High risk of bias

·    Large numbers of dropouts ignored in analysis

Economic evaluation: moderate

·    Fulfilling most critical points of checklist, but did not indicate the chosen perspective.

·    By choosing health care costs and informal care, probably the most important categories were considered, but these is more than a healthcare payer perspective and less than a societal perspective.

Gordon et al.[2],[1] Australia

Clinical data

·    Design: Decision tree model using effectiveness and clinical data from prospective followed cohorts.

·    prospective followed cohorts

·    Setting: 1 university

·    Sample size: 276

·    Recruitment: May 2002-July 2003

·    Data collection: Medical records and self-administered questionnaires (pre-intervention, post-intervention, 6 and 12 months from date of diagnosis), including rehabilitated cases, QALYs and costs

Economic evaluation:

·    Type: CEA, using primary clinical data and modelling (decision tree)

·    Perspective: Societal

·    Cost year & monetary unit: 2004, AU$

·    Length of evaluation: 1-year

Funding: PhD scholarship from the National Breast Cancer Foundation and Women in Super

·    Cancer type: Breast cancer patients

·    Eligibility criteria: Women diagnosed with primary breast cancer, had unilateral disease, spoke English, had no cognitive problems and were aged 25-74 years

Interventions: DAART (Domiciliary Allied Health and Acute Care Rehabilitation Team: Home-based physiotherapy and education vs STRETCH (Strength Through Recreation Exercise Togetherness Care Health): group-based exercise, education and psychosocial intervention

·     DAART, n=36

·     STRETCH, n=31

·     SC: standard care, n=208

Program duration::

·     DAART: 6 weeks (maximum);

·     STRETCH: 8 weeks

Variables included in CEA

·     Effect variables: Rehabilitated cases, QALYs

·     Intervention costs, direct healthcare costs, costs borne by patients and productivity losses (leisure time, volunteers, …)

Effects:

·     Proportion of rehabilitated cases: similar for STRETCH and DAART but slightly higher in SC (not sign. different)

·     Mean adjusted utility scores for DAART: 0.84 (95%CI: 0.77-0.90), STRETCH: 0.80 (95%CI: 0.73-0.87) and SC: 0.72 (95%CI: 0.70-0.75), sign. different.

Costs:

·     Total costs/participant were $1,038 for STRETCH, $342 for DAART and $189 for SC.

Economic evaluation :

·     Rehabilitated cases: SC is dominant above DAART and STRETCH (i.e. more effective and less costly than the interventions);

·     QALY: ICER for DAART and STRETCH is AU$1,344 and AU$14,478, compared to SC

Sensitivity analysis:

·     One-way S.A.; most values did not influence results

Level of evidence (clinical data): High risk of bias

·    Important baseline differences

·    Two different measures of "effectiveness" were used with each producing different results.

Economic evaluation : moderate

Economic evaluation:

·    Fulfilling most critical points of checklist

Jones et al.[3], England

Clinical data

·    Design: Two-arm RCT.

·    Setting: 1 hospice

·    Sample size: 36

·    Recruitment: August 2010-July 2011

·    Data collection: at baseline and after 3-months. These were: Supportive Care Needs Survey Long Form (SCNS-LF59); Kessler Psychological distress Scale (K10); continuity of Care; EQ5D. Service use was collected retrospectively for 3-months from randomization. Societal and demographic data, diagnosis, and disease severity were collected at baseline.

Economic evaluation

·    Type: CUA, using primary clinical data

·          Perspective: Not stated; NHS perspective (at least this threshold is used)

·    Cost year & monetary unit: Not stated (trial year:2010-2011); GBP

·    Length of evaluation: 3-month

Funding: Marie Curie Cancer Care

·    Cancer type: Malignant breast cancer or haematological disease

·    Eligibility criteria: at the end of treatment for first or subsequent recurrence but not cured; with active, progressive, recurrent malignant breast or haematological disease; older than 18 years and meet pre-set referral criteria (i.e. completed treatment, but advanced, progressive disease and recurrence was likely; required symptom management; had rehabilitation needs not responsive to self-management; had psychological, social, financial, emotional, and spiritual needs not met by the present care); and able to reach the hospice by their own or hospice-based transport.

Interventions included four core components, including systematic clinical assessment; goal setting with review and referrals on a case by case basis, according to needs and weekly meetings [2].

·     IG: intervention group: n=20 (allocated n=21).

·     CWL, control waiting list: n=16 (allocated: n=20) ; received usual care (i.e. including ongoing review by oncologists and access to community services including general practitioner (GP), district nurses, social services, and community specialist palliative care), and joined a three-month wait-list for referral to the intervention.

Program duration: ~3-months with flexibility of duration

Variables included in CEA

·     SCNS psychological domain (primary outcome) and as secondary outcomes: other domains; K10, continuity of care, EQ5D (utility and EQ5D VAS)

·     QALY

·     Healthcare utilization (including intervention) & cost

Effects:

·     IG had greater QALY gains than CWL (mean difference 0.05 QALY, 95% CI 0.000-0.112).

·     Primary outcome and other secondary outcomes were significant different at 3-month (e.g. IG had sign. lower needs for support on the psychological subscale of the SCNS than CWL (adjusted difference -16,8 points)). Other significant outcomes included the physical and patient care subscales of the SCNS and the self-reported health state.

·     Other secondary measures all favoured better outcomes in the IG, but without significant differences.

Costs:

·     IG had higher cost than CWL

Economic evaluation :

·     ICER of £ 19,391 per QALY gained. At a WTP of £20,000 or £30,000, the intervention is expected to be cost-effective in 55.4% or 73.3% of simulations, respectively

Sensitivity analysis:

·     No on-way sensitivity analysis

·     PSA using Monte Carlo sampling techniques

Level of evidence (clinical data): high risk of bias

·    Intervention delivered in one single hospice

·    Subjects were not blind about treatment allocation

Economic evaluation: moderate

·    Technical spoken was the analysis in line with the Drummond checklist, except for one-way S.A. but conducted PSA.

·    Negative point: Chosen perspective (i.e. NHS perspective) in line with national guidelines, but a wider perspective might be more appropriated for this population (i.e. only 50% were retired, the others were either employed (~25%), too sick to work or others).

Mourgues et al.[4], France

Clinical data

·    Design: Two-arm, multicenter RCT, stratified by menopausal status.

·    Setting: 1 university hospital and 2 private hospitals

·    Sample size: Economic evaluation, n=90; Trial: n=232 [5]

·    Recruitment: March 2008-October 2010

·    Data collection: at baseline, 6 and 12 months. Women's activities by calculating separately the total hourly volume of overall activities and occupational and non-occupational activities (i.e. primary outcome). Daily abilities (= secondary outcome).

Economic evaluation

·    Type: CEA; using primary clinical data

·    Perspective: Societal perspective

·    Cost year & monetary unit: Not stated; €

·    Length of evaluation: 1-year

Funding: French association of thermal centers, the city of Clermont-Ferrand, the regional council of Auvergne and the association "Ligue contre le Cancer"

·    Cancer type: Complete breast cancer remission

·    Eligibility criteria: women in complete breast cancer remission without contraindication for physical activities or cognitive disorders and a body mass index between 18.5 and 40 kg/m2

Interventions: IG underwent spa treatment (i.e. two week multicomponent programme composed of interventions such as physiotherapy, nutritional advice, thermal water treatment, daily 2-h physical activity, running and basic dietary follow-up over a period of 15 days) combined with consultation with dietician every 6 months;

·    IG, intervention group, n=42 for CEA (trial n=117) [5]

·    SC, standard care & consultations with the dietician every 6 months, n=48 for CEA (trial n=115) [5]

Program duration:

·     2-week spa treatment & consultation with dietician every 6 months

Variables included in CEA

·     Overall activities, occupant and non-occupant activities (and as considered as an effect, productivity losses for absence from paid and unpaid work was not considered)

·     Intervention costs and direct healthcare costs

·     Indirect medical costs comprised out-of-pocket expenses associated with the disease and daily allowances.

Effects:

·     IG had greater resumption of overall activities during the first 12-month period vs SC (p=0.025).

·     There was an interaction effect (p=0.04) with regard to the resumption of occupational activities: more women in IG tended to return to work.

·     Positive effect in the IG on the women's ability to perform occupation activities 12 months after the beginning of the study (p=0.0014), and on their ability to perform family activities (p=0.033).

Costs:

·     Not stated

Economic evaluation :

·     Overall activities: At T6 the thermal treatment was expensive and not cost-efficient. At T12, the intervention was more expensive but also more effective.

·     Occupational activities: At T6, the thermal treatment was too expensive for the moderate increase in effectiveness, whereas at T12 the intervention was slightly expensive but much more effective and therefore cost-efficient.

Sensitivity analysis: Not stated

Level of evidence (clinical data): high risk of bias

·    Subjects were not blind about treatment allocation

·    Many eligible patients rejected to participate in trial

Economic evaluation: moderate

·    Fulfilling most critical points of checklist, except for S.A.

·    No cost result presented and CEA results only presented in figures

Round et al.[6], England [3]

Clinical data, see Jones et al.[3]), and using modelling for extrapolation treatment costs and benefits beyond the initial 3-month follow-up period in S.A.

Economic evaluation

·    Type: CUA, using primary clinical data & modelling

·    Perspective: NHS perspective & a personal social services perspective

·    Cost year & monetary unit: Not stated (~2010-2011); GBP

·    Length of evaluation: 3-month (trial), and S.A. assuming that the benefit of treatment being maintained over three, six and nine months beyond completion of the follow-up

Funding: Marie Curie Cancer Care

·    Cancer type: see Jones et al.[3]

·    Eligibility criteria: see Jones et al.[3]

Interventions: see Jones et al.[3].

Program duration: see Jones et al.[3]

Variables included in CEA

·     QALY

·     Intervention costs and direct healthcare costs

Effects:

·     At 3-months (i.e. trial period) the mean differences in QALYs was 0.052 (95%CI: 0.040-0.063)

Costs:

·     At 3-months (i.e. trial period), the expected mean differences in costs in the base-case analysis was £735 (95%CI: £221 to £1,271)

Economic evaluation :

·     ICER of the mean incremental values is £ 14,231 per QALY gained

Sensitivity analysis:

·     One-way S.A. and PSA.

·     The results of the analysis are sensitive to the method used to estimate QALYs;

·     ‘The longer treatment benefit is maintained, the more likely it becomes that the intervention represents a cost-effective use of resources’

Level of evidence (clinical data): high risk of bias (for details see Jones et al. [3])

Economic evaluation: good

·    Fulfilling most critical points of checklist and presenting detailed results

·    Negative point: Chosen perspective (i.e. NHS perspective) in line with national guidelines, but a wider perspective might be more appropriated for this population, only 50% were retired, the others were either employed (~25%), too sick to work or others.

Mewes et al. [7], the Netherlands

Clinical data:

·    Design: Markov model consisting of four health states: "menopausal symptoms", "reduction in menopausal symptoms", "recurrence" and "death", using effectiveness and clinical data came from a 4-arm RCT of Duijts et al.[8, 9], n=420 randomly allocated using computerized block randomization [9]

·    Setting/sample size: Hypothetical cohort of 1,000 women of 48 years. Trial (multi-center)

·    Recruitment: N.A.

·    Data input: Effectiveness data mainly based on RCT published by Duijts et al.[9], but extrapolated up to 5 years

Economic evaluation

·    Type: CEA; using model

·    Perspective: Dutch healthcare system perspective

·    Cost year & monetary unit: Not stated; €

·    Length of evaluation: Base-case: 6-month; S.A.: 1.5, 3 and 5 years, discounting effects with 1.5% and costs with 4% according to Dutch guidelines

Funding: Alpe d'Huzes, a foundation that is part of the Dutch Cancer Society

·    Cancer type: Breast cancer patients experience (severe) menopausal symptoms after an early onset of menopause caused by cancer treatment

·    Eligibility criteria: Hypothetical cohort of 1,000 patients with a starting age of 48 years and starting in the Markov health state “menopausal symptoms”

Interventions: Comparing cognitive behavioural therapy (CBT) vs physical exercise (PE)[4].

In the original trial[10], sample size per arm was:

·     CBT, n=109

·     PE, n=104

·     CWL: control waiting list: n=103.

Program duration:

·     CBT intervention involved six weekly groups sessions of 90 min each.[9]

·     PE intervention consisted of a 12-week home-based exercise program, individually tailored during an intake session with a physiotherapist. [9]

Variables modelled & included in CEA

·     Deriving QALY, by using SF36 from the trial and converting to EQ5D values

·     Intervention costs, healthcare utilization & cost,

Effects:

·     Total QALY gain was similar across the intervention groups and higher than CWL

Costs:

·     The costs of the interventions were €190 for CBT and €197 for PE

Economic evaluation :

·     ICURs indicate that CBT is likely the most cost-effective treatment, followed by PE as compared to WLC

Sensitivity analysis:

·     One-way S.A. and PSA

·     At a ceiling ratio of €30,000/QALY, the interventions would no longer be considered cost-effective when the duration of treatment effect is 3 or 1.5 years.

Level of evidence (clinical effectiveness data mainly based on Duijts et al., [9]): High risk of bias

Level of evidence (model input data)

·    Model input parameters for base-case and sensitivity analysis are given, including distribution and the used source.

Economic evaluation: moderate

·    Fulfilling most critical points of checklist

·    Negative point: Chosen perspective, a wider perspective might be more appropriated for this population (i.e. 21.9% full-time employed and 53.8%part-time employed);

Retel et al. [11], the Netherlands

Clinical data

·    Design: Markov model with three mutually exclusive health states: “complete remission”, “recurrent disease” and “death” using data from two RCT. Data for usual care (SC) were derived from a multi-center RCT comparing intra-arterial and intervenous chemo radiation in advanced head and neck cancer [12] and data for a preventive (swallowing) exercise program (PREPP) were derived from a clinical trial conducted immediately following the former RCT [13]

·    Setting/sample size: Hypothetical cohort of 1,000 patients of 55 years

·    Recruitment: N.A.

·    Data input: Based on the two RCTs (i.e. [12, 13]) and literature

Economic evaluation

·    Type: CUA, using modelling

·    Perspective: Healthcare perspective

·    Cost year & monetary unit: 2008; €

·    Length of evaluation: 1-year time horizon

Funding: Nothing stated

·    Cancer type: Head and neck cancer patients.

·    Eligibility criteria: Hypothetical cohort of patients aged 55 years and starting with treatment

Interventions: Preventive (swallowing) exercise program. In the original trial:

·     PREEP (i.e. intervention group), n=37

·     SC, standard care, n=43

Program duration: Not stated

Variables included in CEA

·     QALYs partly based on trial, literature and expert elicitation

·     Intervention costs and direct healthcare costs

Effects:

·     QALY: 0.77 (PREEP) vs 0.68 (SC)

Costs:

·     Total health care costs (Treatment + preventive exercise) /patient were: €42,271 for PREEP, and €41,986 for SC

Economic evaluation :

·     ICER of PREEP compared to SC: €3,197 per QALY gained

Sensitivity analysis:

·     One-way and two-way S.A.

·     Majority of analyses resulted in an ICER<€20,000 per QALY;

Level of evidence (model input data).

·    Model input parameters for base-case and sensitivity analysis are given, including distribution and the used source.

Economic evaluation: moderate

·    Fulfilling most critical points of checklist

Arving et al. [14], Sweden

Clinical data

·    Design: RCT with three groups; randomization in blocks[15],

·    Setting: 1 university hospital

·    Sample size: n=168

·    Recruitment: December 1997-December 1999

·    Data collection: Demographic and medical data were retrieved from patient files. Health utilities were measured at baseline and at 1, 3,6, 9, 12 and 24 months

Economic evaluation:

·    Type: CUA using primary clinical data;

·    Perspective: British National Health Service perspective:

·    Cost year & monetary unit: 2004; €

·    Length of evaluation: 2-years (no discounting applied)

Funding: Swedish Cancer Society

·    Cancer type: Breast cancer patients starting adjuvant therapy

·    Eligibility criteria: Breast cancer patients starting adjuvant therapy; ability to speak and understand Swedish; no previous cancer; no on-going psychiatric illness

Interventions took place outside the hospital, face-to-face or over the telephone, and started in median 20 days after inclusion. They were similar and used the same techniques such as relaxation, distraction, activity scheduling, and ways to improve communication, methods derived from cognitive behavioural therapy [16]

·     INS: Psychosocial support from a specially trained nurse, n=55

·     IPS: Psychosocial support from a psychologist, n=57

·     SC: Standard care, n=56.

Program duration::

·     INS: 0-16 sessions (median=2); if ≥1 session: mean (median) duration being 172 (106) days.

·     IPS: 0-23 sessions (median=3); if ≥1 session: mean (median) duration being 210 (178) days.

Variables included in analysis:

·     Health utilities using the EORTC QLQ-C30 translated into the EQ-5D

·     Intervention costs (including salary, a direct hospital component and an indirect allocation (i.e. supervision).

·     Healthcare utilization during 2 years using medical records.

Effects:

·     QALY was highest in INP-group (1.59) compared with INS-group (1.52) and SC-group (1.43).

Costs:

·     Costs (intervention+DHC) were €18,670 for INS, €20,419 for IPS and for SC €25,800.

Economic evaluation :

·     INS and IPS were dominant compared to SC (i.e. INS and IPS had a higher effect (i.e. QALY) and lower costs in comparison to SC).

Sensitivity analysis:

·     Several one-way S.A. performed and basecase results confirmed

·     Bootstrapping with 1,000 replications used to estimate 95%CI

Level of evidence (clinical data): High risk of bias

Economic evaluation: good

·    The analysis was in line with the Drummond checklist. Us both bootstrapping & one-way S.A.

·    Negative point: the chosen perspective is too narrow (i.e. healthcare payer) for a patient population where 65% are employed);

Hollingworth et al.[17], England

Clinical data

·    Design: Unblinded, two-arm, parallel RCT, stratified by recruitment site.

·    Setting: community-setting (2 sites)

·    Sample size: 209 analyzed (220 allocated)

·    Recruitment: October 2009-February 2011

·    Data collection: At baseline and 1, 6 and 12-months. These were: Short-form of the Profile of Mood States (POMS), EORTC QLQ-C30; EQ5D; Trent Patient Views of Cancer Services Questionnaires (only at 6-months). Further healthcare utilization via medical records and intervention costs

Economic evaluation

·    Type: CEA and NMB (using £30,000 per QALY); using primary clinical dataWTP using a threshold of £30,000 per QALY

·    Perspective: National Health Service perspective

·    Cost year & monetary unit: 2010-2011, GBP

·    Length of evaluation: 1-year

Funding: National Institute for Health Research, Research for Patient Benefit

·    Cancer type: Patients starting outpatient radiotherapy or chemotherapy.

·    Eligibility criteria:Age ≥18 and less than 85 years; primary solid tumor diagnosis within previous 12 months; outpatient external radiotherapy over a period of ≥2 weeks or outpatient chemotherapy of ≥two cylces; ability to read and communicate in English; not receiving neoadjuvant chemotherapy; and not diagnosed with ductal carcinoma in situ or skin carcinoma

Interventions: During 2^nd week of radiotherapy/2^nd cycle of chemotherapy, patients completed a face-to-face DT&PL meeting with a radiographer/nurse. A second DT&PL meeting could be arranged toward the end of therapy. The DT&PL forms the basis of a therapeutic conversation where concerns are identified and potential solutions are discussed including immediate staff actions (e.g. providing information), patient actions (e.g. using a self-help resource), and referral (e.g. psychological counselling). These action plans were recorded.

·     IG: intervention group (allocated: n=112; included in intent-to-treat, n=106)

·     SC; standard care (allocated: n=108; included in intent-to-treat, n=103).

Program duration:

·     2 meetings

Variables included in CEA

·     EQ5D (i.e. QALY)

·     Intervention costs and direct healthcare costs

Effects:

·     There was no evidence of an intervention effect on the total POMS score at 12-months or over the 12-month follow-up.

·     Also no sign. difference for QALY or any other secondary outcome

Costs:

·     The intervention cost £19 per patient, and it was not offset by lower subsequent hospital, primary care or medication costs

Economic evaluation :

·     NBM was £20,606 for IG and £22,255 for SC, with ∆-915 (95%CI: -2,398-569). The negative difference in net benefits indicates that the intervention was not cost-effective.

Sensitivity analysis:

·     Subgroup analysis

Level of evidence (clinical data): High risk of bias

• Unblinded RCT

Economic evaluation: moderate

·    Fulfilling most critical points of checklist, except for sensitivity analyses. Authors conducted subgroup analyses, but no sensitivity analysis.

·    Negative point: the chosen perspective might be too narrow (i.e. healthcare payer) for this population group (30%-40% are still "in work"). Negative: limited S.A.

Lemieux et al.[18], Canada

Clinical data

·    Design: Blind two-arm RCT, stratified by center and the presence or absence of visceral metastases.

·    Setting: 7 centers (but only 3 of the 7 for the economic evaluation)

·    Sample size: economic analysis using only patients from 3-sites; n=125

·    Recruitment: 1993-1998

·    Data collection: at base line, 4, 8, and 12 months, using psychosocial questionnaires that included the Profile of Mood States and the pain and suffering scales used by Spiegel and Bloom and the EORTC QLQ-C30. Further, information on demographic characteristics and social support.[1]

Economic evaluation

·    Type: CMA (for primary outcome) and CEA for mood and pain; using primary clinical data

·    Perspective: Healthcare system

·    Cost year & monetary unit: 2002-2003; CAN$

·    Length of evaluation: Not stated, ~1-year (i.e. effect is measured at one-year, although length of follow-up is 722 days (IG) and 750 days (SC))

Funding: Canadian Institute of Health Research and the Canadian Breast Cancer Research Alliance.

·    Cancer type: Breast cancer patients.

·    Eligibility criteria: Women who had histologic confirmation of breast cancer at the time of diagnosis, if they had metastases outside of the breast and ipsilateral axilla, and if the treating physician most responsible for a woman’s care gave consent[1]

Interventions: Weekly, 90-minute, therapist-led support group that adhered to principles of supportive-expressive (SE) therapy. Every four to six months, all the women received educational materials about breast cancer and its treatment, as well as about relaxation and nutrition.

·     IG: intervention group, n=43

·     SC: standard care & educational materials, n=82;

Program duration: Attending the group sessions for at least one year, or longer if the sessions continued to be of benefit

Variables included in CEA

·     Survival (primary outcome)

·     Secondary outcomes: psychosocial functioning, mood, pain,

·     Intervention costs and direct healthcare costs

Effects:

·     No significant difference between both groups in survival

·     Statistically significant benefits were found in psychological distress (0.32 for POMS-TMD) and pain (0.40 PAIN-VAS) over the 1st year.

Costs:

·     The control costs were $2,169

·     The mean cost of care per patient was $28,189 and $31,715 in SC and IG, respectively.

Economic evaluation :

·     CMA: Difference between both arms were equal to $3,526 (not significant), and if reducing intervention costs (i.e. $2,169), there was no statistically significant difference in resource costs between IG and SC.

·     CEA: incremental costs are CAN$5,550 and CAN$4,309 for an effect size of change in mood and pain, respectively

Sensitivity analysis:

·     One- way S.A. No change in results.

Level of evidence (clinical data): High risk of bias

Economic evaluation: moderate

·    Analysis was in line with the Drummond checklist.

·    Negative point: the narrow perspective chosen (i.e. healthcare payer)

·    Subset of participants (i.e. 3 of the 7 centres) for economic evaluation

Mandelblatt et al.[19], USA

Clinical data

·    Design: Three-arm RCT, stratified by study site, whether the woman had received chemotherapy, and marital status (married/living as married v other); randomization based on a random number-generated list.

·    Setting: 3 sites

·    Sample size: 388

·    Recruitment: July 1999-June 2002

·    Data collection: At baseline, 2-months (~4 to 6 weeks) after primary treatment; and at 6 and 12 months after intervention, using IES-R and MOS-SF36. Further included the baseline demographic and clinical data, the 2-month   asses if in the IG women had watched the videotape. Further, every 3 months documenting health services. Research staff used weekly logs to record time and resources used to deliver   intervention.

Economic evaluation

·    Type: CEA; using primary clinical data

·    Perspective: Societal perspective

·    Cost year & monetary unit: Not stated (~2002); US$

·    Length of evaluation: 6-month 'because this is the period of immediate transition and by 12 months, most women have adjusted to survivorship'[19]

Funding: National Cancer Institute

·    Cancer type: Breast cancer patients

·    Eligibility criteria: Women who had received surgery for invasive breast cancer of any size or nodal status, and who had no neoadjuvant chemotherapy, high-dose chemotherapy with bone marrow or stem-cell rescue or protracted reconstructive surgery, and who were able to read and write in English

Interventions:: Videotape intervention and printed information (VID) vs psychological educational counselling , videotape and printed information (EDU)

·     VID, n=128

·     EDU, n=135

·     SC, standard care & printed information, n=125.

Program duration:

·     VID: not stated

·     EDU: 2 sessions, the first 80-minutes and the 2nd 2 weeks later by phone, 30-minutes

Variables included in CEA

·     Distress and energy 6 months postintervention, using IES-R and MOS-SF36 vitality scale

·     Intervention costs, healthcare utilization and patients time cost

Effects:

·     EDU was not more effective in increasing energy or decreasing distress than the other arms.

Costs:

·     Intervention costs were $11.30 for SC; $25.85 for VID and $134.47 for EDU

·     No significant differences in health care costs over the 12 months post-intervention period by study arm.

Economic evaluation :

·     EDU was not more effective than the two others, but more expensive, thus dominated by the two others.

·     ICER for VID vs SC was $7,275 per unit of decreased distress and $2.22 per unit improvement in energy, respectively

Sensitivity analysis:

·     One-way S.A.

·     No change of results

Level of evidence (clinical data): High risk of bias

·    Study arms were unbalanced for baseline depression

Economic evaluation: moderate

·    Analysis was in line with the Drummond checklist.

·    States to have used the societal perspective, but unclear if that was indeed the case

Bradley et al.[20], England

Clinical data

·    Design: prospective enriched cohort study. Patients from 3 of 12 hospitals formed IG. Patients from remaining hospitals (i.e.9) formed control group (SC). Matching criteria were: Age, lung function comorbidity and type of surgery

·    Setting: 12 hospitals

·    Sample size: 363

·    Recruitment: Not stated

·    Data collection: Demographic, clinical and healthcare cost data were collected pre-rehabilitation, post-rehabilitation presurgery, 4 weeks post-surgery and at 6 months.

Economic evaluation

·    Type: CBA; using primary clinical data

·    Perspective: Not stated (applied Healthcare payer)

·    Cost year & monetary unit: 2010-2011; GBP

·    Length of evaluation: Not stated (~ 6 months)

Funding: Nothing indicated

·    Cancer type: Patients undergoing curative lung cancer surgery

·    Eligibility criteria: Patient who was considered fit for curative lung cancer surgery by lung cancer multidisciplinary team at regional thoracic unit and following BTS guidelines.

Interventions: Program to optimize physical status, prepare for inpatient journey and support through recovery after surgery. Includes exercise classes, smoking cessation, dietary advice and patient education[5].

·     IG: intervention group, n=58 (only 28 managed to attend the postoperative element)

·     SC: standard care, n=305.

Program duration: Not stated (~ 6 months)

Variables included in CBA

·     Postoperative pulmonary complication; readmission; length of admission, … (expressed in natural units & as costs)

·     Healthcare costs

Effects:

·     Patients in IG had fewer Postoperative pulmonary complications than SC (9 vs 16%, p=0.21) and fewer readmissions (5 vs 14%, p=0.12).

Costs:

·     Total cost/patient in IG was estimated at £1284 compared with £1528 for SC.

Economic evaluation :

·     IG compared to SC results in savings of £244/patient

Sensitivity analysis: Not applied

Level of evidence (clinical data): High risk of bias

·    Significant decrease of participants in the postoperative rehabilitation

Economic evaluation : moderate

·    No health effect measured.

·    Not performing sensitivity analyses

·    Negative point: Narrow perspective (i.e. direct healthcare costs only); a wider perspective might be more appropriated for this population

Björnekl et al.[21], Sweden

Clinical data

·    Design: RCT with two groups; randomization in blocks of four with closed envelops

·    Setting: 1 hospital

·    Sample size: 382

·    Recruitment: April 2002-November 2007

·    Data collection: Self-reported questionnaires at baseline (i.e. after randomization but before intervention), 2, 6 and 12-months after intervention. Family situation, occupation, sick leave and healthcare utilization

Economic evaluation

·    Type: CBA; using primary clinical data

·    Perspective: Societal

·    Cost year & monetary unit: Not stated (trial period); SEK

·    Length of evaluation: 1-year

Funding: Country Council of Västmanland, the Swedisch Social Insurance Agency, the Västmanland Research Fund against Cancer and the National Federation of Cancer and Traffic Injury

·    Cancer type: Breast cancer.

·    Eligibility criteria: Newly diagnosed primary breast cancer, no previous malignancy, the physical and mental capability to participate in group interventions and to fill in questionnaires and an expected survival time of more than 12 months.

Intervention Information-based support program supplemented with relaxation, qi-gong and liberating dance taking place within 4-months of ending treatment; comprising a 7-day stay at a resort, where participants take part in the support program, followed by a 4-day follow-up 2-months after the initial visit.

·     IG: Intervention group, n=191

·     SC: Standard care, n=191

Program duration: ~2.5 months

Variables included in CEA

·     Sick leave of patient (number of days & expressed as costs (i.e. productivity losses))

·     Health care utilization (expressed in natural units & as costs)

Effects:

·     No sign. difference between the groups, neither for sick leave, nor the number of visits to medical specialists at any time after the intervention period.

Costs:

·     At all points in time higher costs for sick leave and consumption of health services for IG than SC and sign. differences between groups after 12-months. Adding the cost of the intervention made the cost for the IG statistically significantly higher at all times of measurement.

Economic evaluation :

·     SC is dominant compared to IG. No sign. difference in effects between groups and higher costs for IG.

Sensitivity analysis: N.A.

Level of evidence (clinical data): High risk of bias

·    Authors highlighted the fact that the intervention might have influenced the patients' thoughts and feelings and created a need for sick leave to handle and cope with their anxiety

Economic evaluation: moderate

·    No health effect measured.

·    No sensitivity analysis applied or similar. For all other points mainly in line with the Drummond checklist.

Tamminga et al.[22], the Netherlands

Clinical data

·    Design: Two-arm RCT, randomization using computerized randomization program ALEA; stratified by return-to-work, age (<50 or ≥50 years) and cancer diagnosis. Patients, nurses and researchers are not blind to group assignment.

·    Setting: 6 hospitals

·    Sample size: 121 analyzed (133 allocated)

·    Recruitment: May 2009-December 2010

·    Data collection: At baseline, 6 and 12-month. Socio-demographic factors and prognostic factors for time until return-to-work were assessed at baseline only. Oucome measures (e.g. return-to-work and qol) and cancer treatments were assessed at all-time points. Intervention details were collected from nurses.

Economic evaluation

·    Type: CMA (no CEA as no sign. differences between groups on outcomes measured); using primary clinical data

·    Perspective: Societal

·    Cost year & monetary unit: Not stated; €

·    Length of evaluation: for economic evaluation, only first year follow-up

Funding: Stichting Instituut Gak

·    Cancer type: Breast and gynaecological cancer

·    Eligibility criteria: Cancer patients between 18 and 60 years of age who had been treated with curative intent, had paid work, who were on sick leave; were able to speak, read and write Dutch, had no severe mental disorder or other severe comorbidity. Treatment with curative intent was defined as an expected 1-year survival rate of approximately 80%. We excluded patients who were not sufficiently able to speak, read, or write Dutch, had a severe mental disorder or other severe comorbidity, and for whom the primary diagnosis of cancer had been made more than two months previously.

Interventions: included: 1) 4 meetings of 15 minutes each as part of the normal consulting hour to start early vocational rehabilitation carried out by an oncology nurse, social worker or nurse practitioner; 2) one meeting with the participant, the occupational physician, and the supervisor to make a return-to-work plan, and 3) three letters send to the occupational physician to enhance communication; two will be from the treating physician and one from the nurse.

·     IG, intervention group, n=61 analyzed (65 allocated)

·     SC, standard care, n=60 analyzed (68 allocated).

Program duration: Not stated

Variables included in CEA

·     Rate of return-to work at one year of follow-up

·     Number of days between the first day of sick leave and the first day at work sustained for at least 4 weeks.

·     Qol using SF-36, including all subscales and VAS.

·     Work ability using the first question of WAI.

·     Impaired work functioning using WLQ

·     Intervention costs

·     Lost productivity costs and work adjustments costs

·     No healthcare utilization

Effects:

·     Study failed to show any significant differences between groups on return-to-work outcomes and qol.

Costs:

·     Intervention costs were €119/patient in IG

·     The mean lost productivity cost according to the human capital approach was €41,393 in IG and €38,968 in SC. The mean productivity cost according to the friction costs approach was €14,030 in IG and €13,529 in SC.

·     The mean work accommodations cost was €2,975 and €3,025 in IG and SC, respectively.

·     These costs did not differ statistically between groups

Economic evaluation :

·     No statistical significant effect and costs between groups.

Sensitivity analysis: Not applied

Level of evidence (clinical data): High risk of bias

·    No bliniding of participants, nurses and patients.

·    Potential “contamination” of the SC.

·    Treating nurses not convinced of their competence & occupational physician and employer were not interested.

Economic evaluation: poor

·    No health effect measured.

·    Did not conduct scenario analysis

·    Do not present ICERs although that was the aim (mainly because they failed to show any differences between the two groups).

·    Stated to take a societal perspective, no information on healthcare resources were collected, only intervention costs and productivity losses