Oncogeriatrics health economic simulation

Introduction

This repository accompanies a scientific paper currently under peer-review. McKenzie et al., undertook a synthetic model-based health economic evaluation of geriatric assessment prior to cancer treatment. A model developed using the Python programming language has been packaged to share with the research community.

Description

This Python model has been packaged using Poetry and can be used as follows. If you have Poetry already installed or wish to use it for this project, follow the steps below: -

Installation with Poetry

First, clone the repository into a suitable folder on your machine, following the instructions here.

Second, initialise the poetry shell by executing the following command in your terminal: -

poetry shell

Thirdly, install the dependencies: -

poetry install

Installation without Poetry

You can of course install and run this model without Poetry, provided the following dependencies are considered: -

• python = ">= 3.7.3, < 3.11"
• numpy = "^1.21.2"
• PyYAML = "^5.4.1"
• tabulate = "^0.8.9"
• pandas = "^1.3.3"
• scipy = "^1.7.1"

Running

For convenience a Makefile has been created to enable easing running of the model once installed. To run execute the following terminal command: -

make run

Without using poetry, execute: -

python model.py

Model modification

We encourage other researchers to reuse this model and have therefore made it easy to adjust basic parameters with little coding knowledge. For more advanced modification, Python programming experience is required, but this model can be used as a foundation.

To adjust basic assumptions of the model, open the file assumptions.yaml. This is a YAML Ain't Markup Language (YAML) file and is human and machine readable.

Basic

Parameters	Description	Default
simulations	The number of Monte Carlo simulations to undertake	5000
progress	Whether to show a progress bar or not in the terminal output	True
treatment-distributions	The distributions of different treatments in the following order: - other care chemotherapy only surgery only radiotherapy only radiotherapy only chemotherapy and radiotherapy surgery and chemotherapy surgery and radiotherapy surgery, radiotherapy and chemotherapy	[0.335175413, 0.083228036, 0.209740598, 0.147376245, 0.049030028, 0.074862494, 0.072866805, 0.027720381] derived from NHS Cancer Data, assumed for a 77-year-old patient

Implementation

Parameters	Description	Default
tablet-based-assessments	Patients have the option of completing a geriatric assessment on a tablet device	False
face-to-face-assessments-nurse	Patients are being seen face-to-face by a nurse	True
face-to-face-assessments-consultant	Patients are being seen face-to-face by a Consultant (Attending) Physician	True
face-to-face-assessments-registrar	Patients are being seen face-to-face by a Specialist Registrar (Resident) Physician	False
telephone-assessments	A nurse is undertaking telephone assessments	False
remote-assessments-where-possible	Where possible, remote patient-reported assessements are being undertaken	False
ga-changing-management-at-mdt-level	The results from the geriatric assessment are changing management at the cancer multi-disciplinary team (tumour board) level	True
only-undergoing-chemotherapy	Patients are only undergoing geriatric assessment before chemotherapy	False
only-undergoing-surgery	Patients are only undergoing geriatric assessment before surgery	False

Clinical effectiveness of Geriatric Assessment

Parameters	Description	Default
reduced-los-effect	Geriatric assessment reduces length of stay (relative risk)	1
reduced-chemotherapy-toxicity-effect	Geriatric assessment reduces chemotherapy toxicity	True
reduced-er-visits-effect	Geriatric assessment reduces emergency department (room) visits (relative risk)	1
reduced-itu-admissions-effect	Geriatric assessment reduces intenstive therapy unit admissions (relative risk)	1
reduced-surgical-complications-effect	Geriatric assessment reduces post-operative complications	True
reduced-post-surgical-readmissions-effect	Geriatric assessment reduces post-operative complications (relative risk)	1

Surgical parameters

Parameters	Description	Default
bed-days-alpha	The α parameter of a Gamma distribution for postoperative bed days	2.14
bed-days-beta	The β parameter of a Gamma distribution for postoperative bed days	3.04
requiring-itu	The percentage requiring Intensive Care Unit admission postoperatively	0.097
readmissions-alpha	The α parameter of a Beta distribution for postoperative readmissions	3.6
readmissions-beta	The β parameter of a Beta distribution for postoperative readmissions	31.5

General parameters

Parameters	Description	Default
er-visits-alpha	The α parameter of a Beta distribution for emergency department (room) visits	0.78
er-visits-beta	The β parameter of a Gamma distribution for emergency department (room) visits	6.31
initial-qaly-alpha	The α parameter of a Beta distribution for the initial quality-adjusted life year before treatment	37.79
initial-qaly-beta	The β parameter of a Beta distribution for the initial quality-adjusted life year before treatment	13.93
10-year-survival-probabilities	The 10-year survival probabilties of 77-year-old adults with cancer -/+ postoperative complications	[[ 0.967711, 0.612698, 0.52797, 0.473091, 0.438158, 0.413269, 0.3908, 0.370723, 0.354533, 0.343723, 0.339788 ], [ 0.967711, 0.571380518, 0.464723943, 0.395256173, 0.356781234, 0.328020278, 0.301236405, 0.277092946, 0.259023145, 0.243354875, 0.230547149 ]]
nice-recommended-yearly-discount	The National Institute of Health and Care Excellence recommended yearly discount in quality-adjusted life years	0.035
chemotherapy-qaly-decrement-alpha	The α parameter of a Beta distribution for the decrement in quality-adjusted life years during and following chemotherapy (lasts on year)	77.05
chemotherapy-qaly-decrement-beta	The β parameter of a Beta distribution for the decrement in quality-adjusted life years during and following chemotherapy (lasts on year)	163.73

Unit costs

The unit costs can be found in the utilisation_costs.yaml file. These are again editable to account for cost changes and international differences.

Pretreatment costs

Parameters	Description	Default (£)
ga-using-tablet-technology	The cost per patient to use technology to assist geriatric assessment	2
ga-using-tablet-staff	The cost of staff per patient to use technology to assist geriatric assessment	68.78
ga-using-consultant	The cost of a Consultant (attending) Physician to undertake a 30-minute component of geriatric assessment	141.18
ga-using-registrar	The cost of a Specialist Registrar (Resident) Physician to undertake a 30-minute component of geriatric assessment	169.23
ga-using-nurse-f2f	The cost of a Band 6 nurse to undertake a 60-minute component of geriatric assessment	117.91
ga-using-telephone-nurse-led	The cost of a Band 6 nurse to undertake a 30-minute geriatric assessment over the phone	117.91
dietician	The cost of a dietician to undertake a 30-minute assessment	58.96
social-worker	The cost of a social worker to undertake a 30-minute assessment	32.51
occupational-therapy	The cost of a occupational therapist to undertake a 30-minute assessment	60.23
physiotherapist	The cost of a physiotherapist to undertake a 30-minute assessment	60.23
falls-clinic	The cost of a falls clinic attendance	747.07
outpatient-physician	The cost of a 30-minute appointment with an outpatient Consultant (Attending) Physician	141.18
cbt-course	The cost of a course of Cognitive Behavioural Therapy per patient	1053.40

Posttreatment costs

Parameters	Description	Default (£)
excess-bed-day	The cost of an excess bed day following surgery	366.01
hdu-or-itu-admission	The cost of a High Dependency Unit or Intensive Care Unit admission	2160.21
chemotherapy-toxicity-short-stay	The cost of a short (1 day) admission for chemotherapy toxicity	614.75
chemotherapy-toxicity-long-stay	The cost of a long (5 days +) admission for chemotherapy toxicity	3437.29
er-visits	The cost of an emergency department (room) visit	169.92
surgical-readmission	The cost of a readmission following surgery	3522.70

Unmet needs

Unmet need is difficult to model economically, but the differences in the identification of unmet needs between standard care and introducing geriatric assessment are contained within the parameters.yaml file. The key values to modify are the mean, upper and lower limits of each parameters for the relevant arm (standard care versus standard care and geriatric assessment).

License

This code is MIT licensed (see license.md) and we encourage other researchers to reuse and advance this model for the benefit of oncogeriatric research. We would be most grateful if you could kindly cite our final article when published if you have used this model in your research.