Model inputs
QALY es(mates are derived from the literature (Lidgren et al, 2007). The following health states were iden(fied:
Transi(on probabili(es (λ1, λ2, λ3) were es(mated from Adjuvant!
and IBTR!. The transi(on probabili(es were adapted to obtain cumula(ve distribu(on func(ons for the risk using three criteria: (1) age, (2) tumor size and (3) lymph node involvement. Es(mates for the (me of a recurrence were derived from Engel et al, 2003a and 2003b.
Assump/ons for the effect of follow-‐up programs
• Surgeon and nurse-‐prac((oner follow up perform equal in terms
of detec(ng recurrences;
• Telephone consulta(on is assumed to be not-‐effec(ve for
detec(on of recurrences (probability = 0).
• The effect of adjuvant therapy is modeled in the risk rates;
• Regarding the probability of detec(ng a recurrence we used the
following assump(ons:
OPTIMIZATION OF FOLLOW-‐UP SCENARIOS FOLLOWING BREAST CANCER
Maarten IJzerman, PhD
1
, Erwin Hans, PhD
2
, Sabine Siesling, PhD
1
and Joost Klaase, MD, PhD
3
(1) Dept. Health Technology & Services Research, University of Twente, (2) Dept. Opera(onal Methods and Logis(cs, University of Twente and (3) Dept. Surgical Oncology, Medisch Spectrum Twente
Objec/ves
About one in every eight women develops breast cancer.
In the Netherlands, 11,000 new cases are registered every year and about 3500 women die of breast cancer. According to the guidelines
(www.oncoline.nl for the Netherlands) most pa(ents are currently
assigned the same follow up, i.e. five years long, two consults per year. It was inves(gated whether a less intensive follow-‐up scheme may be more appropriate.
Conclusion
• In general, we can conclude that young pa(ents (<50)
require a more intensive follow-‐up than older pa(ents (>70). Older pa(ents have a lower life expectancy, and therefore there are less QALYs to be gained and the effec(veness of follow-‐up is lower.
• Pa(ents with (very) unfavorable tumor characteris(cs
s(ll benefit from follow-‐up.
• The number of consults can be reduced drama(cally by
switching to an individualized follow-‐up.
References
1. Robertson C: The clinical effec(veness and cost-‐effec(veness of different surveillance mammography regimens afer the treatment for primary breast cancer: systema(c reviews registry database analyses and economic evalua(on. Health Technol Assess 2011;15(34) 2. Lidgren M., Wilking N., Jönsson B., Rehnberg C. (2007). "Health related quality of life in
different states of breast cancer." Quality of Life Research 16(6): 1073-‐1081
3. Kimman ML: Economic evalua(on of four follow-‐up strategies afer cura(ve treatment for breast cancer: results of an RCT. Eur J Cancer. 2011 May;47(8):1175–1185.
4. Kimman ML, Dellaert BGC, Boersma LJ, Lambin P, Dirksen CD. Follow-‐up afer treatment for breast cancer: one strategy fits all? An inves(ga(on of pa(ent preferences using a discrete choice experiment. Acta Oncol. 2010 Apr.;49(3):328–337.
5. Engel J., Eckel R., Kerr J., Schmidt M., Fürstenberger G., Richter R., Sauer H., Senn H.-‐J., Hölzel D. (2003a). "The process of metastasisa(on for breast cancer." European Journal of Cancer 39(12): 1794-‐1806
6. Engel J., Eckel R., Aydemir U., Aydemir S., Kerr J., Schlesinger-‐Raab A., Dirschedl P., Hölzel D. (2003b) "Determinants and prognoses of locoregional and distant progression in breast cancer." Interna(onal Journal of Radia(on Oncology Biol. Phys. 55(5):1186-‐1195.
Presented at
October 23
th, 6:30PM
Discrete-‐Event Simula/on
Discrete-‐event simula(on was used to calculate the most efficient use of workforce alloca(on. Tecnoma(x Plant Simula(on sofware
was used for simula(on (www.plm.automa(on.siemens.com).
State-‐transi/on model
Simula/on approach
The simula(on starts with the crea(on of pa(ent groups of 1000 pa(ents each. For each pa(ent group 300 runs were simulated. Assuming three criteria, 120 different pa(ent groups were iden(fied. The simula(on starts with the genera(on of disease processes for each individual pa(ent. Simula(on con(nues un(l all pa(ents have died.
Background
Kimman et al (Eur. J. Cancer, 2011)
One-‐year cost-‐effec(veness of four follow-‐up scenarios:
hospital follow-‐up; (2) nurse-‐led telephone follow-‐up; (3) hospital follow-‐up plus educa(onal group program (EGP); and (4) nurse-‐led telephone follow-‐up plus EGP. Nurse-‐led telephone follow-‐up plus EGP seems an appropriate and cost-‐effec(ve alterna(ve to hospital follow-‐up for breast cancer pa(ents during their first year.
Kimman et al (Acta Oncol, 2010)
The medical specialist was the most preferred to perform the follow-‐up, but a combina(on of the medical specialist and breast care nurse alterna(ng was also acceptable to pa(ents. Face-‐to-‐face contact was strongly preferred to telephone contact. Follow-‐up visits every three months were preferred over visits every four, six, or 12 months.
Robertson et al (HTA, 2011)
Combining ini(a(on, frequency and dura(on of surveillance mammography resulted in 54 differing surveillance regimens for women afer BCS and 56 for women following mastectomy. The studies included in the clinical effec(veness review suggest surveillance mammography offers a survival benefit.
create pa(ents
(1000 pa(ents, 300 runs)
Generate disease
process per pa(ent Increase (me with one year
Set age, tumor size and lymph node status according to pa(ent group
Start simula(on
Repeat un(l all pa(ents have died
Record consulta(ons and QALYs
Determine dead from other causes
(λ5)
Move pa(ent to other health states because of BC events
Every year
Generate local recurrence and (me
of event (λ1 and λ3)
Generate second primary tumor and
(me of event (λ2)
Generate primary metastasis and (me
of event (λ4)
Generate LR of second primary tumor and
(me of event (λ3)
[if second primary]
Metastasis risk (λ4) Second primary tumor (λ2) Local recurrence risk (λ1)
age tumour size lymph node involvement
Simula/on objec/ve
The main objec(ve of the simula(on was the op(miza(on of capacity planning from a hospital perspec(ve, taking into account the heterogeneity in case mix. We assumed the number of consulta(ons to be the op(miza(on criterion and have set this at 40 consults throughout the follow-‐up period.
Simula/on results
Typical simula(on for one group of pa(ents. Assuming a threshold Of 40 consulta(ons per QALY, the most intensive follow-‐up is preferred.
The figure presents the recommended follow-‐up scenarios for all 120 pa(ent groups.
Red: most intense follow-‐up is recommended in these pa(ent groups. Green: least intensive follow-‐up may be recommended in these pa(ent groups.
Es/mated consequences
Implementa(on of the follow-‐up recommenda(ons may be beneficial for capacity planning. We assumed the face-‐to-‐face interview by a surgeon to take 10 minutes. A face-‐to-‐face interview by a NP would take 20 minutes.
Address for correspondence:
University of Twente, Dept. Health Technology & Services Research
PO Box 217
7500 AE Enschede
www.utwente.nl/mb/htsr
The Netherlands
m.j.ijzerman@utwente.nl
Impact on capacity planning
Several follow-‐up scenarios may be evaluated on the ability to reduce capacity for breast cancer follow-‐up. Present guidelines result in nearly 80,000 hospital visits. Changing exis(ng Oncoline guidelines may save up to 22% of required capacity. However, an individualized approach accep(ng max. 40 visits per QALY gained would lead to about 70% reduc(on in required capacity.
