Melanoma is a type of skin cancer that develops from skin cells called melanocytes. Around 15,400 people are diagnosed with melanoma in the UK yearly, and the rate of new diagnoses of melanoma in the UK has risen faster than other cancers. Advanced melanoma means the melanoma has spread from its origin to another part of the body. In the past few years, more treatment options have become available for this type of cancer, such as immunotherapies, which are treatments that use the immune system to recognise and kill cancer. New treatments have led to more patients living longer.
Because people now live longer with advanced melanoma, they are sometimes thought of as “survivors”; and the state of surviving beyond a certain timepoint “survivorship”. Nonetheless, there is no standard definition of “survivorship” in advanced melanoma. It might include people surviving several years after their diagnosis or the end of their treatment.
The project will use the UK Clinical Practice Research Datalink (CPRD) database to investigate melanoma “survivors”, and determine their characteristics (e.g., age and gender) and health status. Second, we will also investigate and describe patients who use immunotherapies. Finally, we will determine how often and to what level survivors of advanced melanoma and treated with immunotherapies in the past use healthcare in England, such as doctors’ surgeries, hospitals, and specialist clinics.
Survival of advanced melanoma patients has considerably improved thanks to new therapeutics tools such as immunotherapies, with patients surviving beyond 5 years of diagnosis being considered “survivors”. Currently, there is no clear understanding of the characteristics of survivors, their healthcare resource use (HCRU) and the associated burden to the English health system.
This retrospective cohort study will use population-level data from England. The Clinical Practice Research Datalink (CPRD) AURUM dataset linked with Hospital Episode Statistics (HES) Accident and Emergency, Admitted Patient Care, Outpatient data, and National Cancer Registration and Analysis Service (NCRAS) Cancer Registration Data and Systemic Anti-Cancer Treatment (SACT) data will be utilized.
In advanced melanoma survivors, this project aims to understand: (1) The patient characteristics, treatment patterns and survival trends; (2) Their HCRU compared with the general population; and (3) The type of services after survivorship such as outpatient clinics attended.
Exposures will be survival for five years post diagnosis of advanced melanoma and treatment with immunotherapies (Ipilimumab, Pembrolizumab, Nivolumab, Talimogene Laherparepvec). Outcomes will be survival measures, immunotherapy treatment sequences, and HCRU.
The non-parametric Kaplan-Meier method will be used to describe overall survival (OS) at one to five years. Landmark survival analysis and extended modelling with time-varying covariate using Cox regression will be employed to assess conditional survival. Within each treatment regimen, we will describe the specific drugs, duration, reason for discontinuation, number of cycles, dosing, duration and route of administration. The number and proportion of patients experiencing each HCRU event (GP/nurse visit, medication, outpatient specialist visit, inpatient admission, etc.) will be presented per year. The rate of events per patient per month (and the 95% Confidence Interval) will be calculated.
The study findings will benefit patients in England by filling a knowledge gap on treatment and outcomes of patients who have survived malignant melanoma and their HCRU.
The following outcomes will be ascertained for patients with metastatic melanoma
• Survival measures
• Treatment sequences
In a subpopulation of patients with metastatic melanoma that will be considered as survivors (still alive after 5 years of diagnosis; defined in Section L: Definition of Study Population), the following outcomes will be measured:
• HCRU
The key study definitions, in relation to outcomes to be measured, are:
• Index date: An index date will be defined for each sample as the earliest date at which a patient meets the eligibility criteria for the respective sample. Index dates will be defined separately for each study sample (described in Section L: Definition of the Study Population). Non-melanoma comparators will be assigned an specific index date according to matching criteria (described in Section M: Selection of comparison group(s) or controls).
• Baseline (pre-index) period: This period will span from 1 January 2009, the date of birth, or an individuals’ first CPRD record date (whichever came last); and up to the day before the index date. A minimum baseline period of 12 months prior to the index date is required.
• Follow-up periods: These periods will start at the index date for each sample and end at death, disenrollment from the database, or end of the study period, whichever comes first.
Survival measures
The following survival measures will be defined using time (months) since index date. They will also be measured at a 5-year landmark timepoint (conditional survival – conditional on having survived a specific amount of time):
o Overall survival (OS) (primary objective)
o Progression-free survival (PFS) (exploratory objective)
o Progression on second-line treatment (exploratory objective)
o Post-progression survival (exploratory objective)
o Treatment-free survival (TFS) (exploratory objective).
A sensitivity analysis at a 3-year landmark timepoint will also be conducted.
Table 1 provides working definitions for each measure, in addition to the variables utilised to apply each definition.
The use of National Cancer Registration and Analysis Service (NCRAS)/ Systemic Anti-Cancer Therapy Dataset (SACT) variables is specified where relevant, and highlighted in pink throughout this protocol for clarity.
Table 1. Survival Measures – Operational Definition and Variables Used (see attached)
Treatment sequences
Using SACT data, treatments administered in each line will be described to understand treatment sequencing. As treatment line is no longer available in SACT,2 we will leverage our experience from previous studies in which we have identified alternative variables in this dataset that can be used to define lines of therapy (LOT). The proportion of patients receiving treatments in each line will be described.
The following variables from the SACT dataset will be ascertained for each patient:
o Name of drug(s) (SACT variable ‘Analysis_Group’)
o Dose of each drug (SACT variable ‘Actual_Dose_Per_Administration’)
o Route of administration of each drug (SACT variable ‘Administration_Route’)
o Regimen duration (SACT variable ‘Start_Date_of_Regimen’ ‘)
Line of therapy will be defined according to the following criteria:
LOT start date: The LOT start date will be defined as the date of the first administration of the first cycle of a regimen, defined as the earliest date of: start date of regimen, (SACT variable ‘Start_Date_of_Regimen’ in SACT), cycle start date (SACT variables ‘start_date_of_cycle’ and ‘cycle_number’), or drug administration date (SACT variable ‘administration_date’).
LOT end date3: The LOT end date will be defined as the earliest date of:
• 180 days after the latest of:
o Cycle start date (SACT variables ‘start_date_of_cycle’ and ‘cycle_number’) without starting a new regimen (SACT variable ‘Start_Date_of_Regimen’), where a regimen contains several cycles which are uninterrupted periods of time in which the treatments are administered, or
o Drug administration date (SACT variable ‘administration_date’), when different to start date of cycle
• The day before the start of a new regimen (SACT variable ‘Start_Date_of_Regimen’), or
• Death (CPRD-HES-ONS: Death record, NCRAS: variable “vitalstatus” and “vitalstatusdate”), or
• End of study period.
LOT will be defined as 1-5+. The grouping may be subject to change depending on sample size in the higher LOT groups (e.g., 1-3+).
HCRU
The following outcomes will be ascertained during the follow-up period for all patients with advanced melanoma treated with immunotherapies. Based on these outcomes, rates for each HCRU category will be derived.
Primary care HCRU
o General practitioner (GP) consultations/nurse visits; identified using CPRD Aurum Consultation files
o GP referrals; involving both inbound and outbound patient referrals to or from external care centres; identified using CPRD Aurum Referral file. Note CPRD Aurum variable “referral target organisation”, which would be of interest for this study, is not currently populated and thus the type of organisation the patient is referred to is not available as of October 2021. Number of referrals will however be recorded and outpatient visits and procedures in hospitals and clinics will be captured (see below).
o GP prescriptions; identified using CPRD Aurum drug issue files. This includes GP-issued prescriptions and only a very small proportion of secondary care prescriptions. Note the inclusion of secondary care prescriptions is largely incomplete and varies by individual practice approach.
Secondary Care HCRU
Inpatient HCRU
o Total Number of hospital admissions: Total number of admissions (including all admission diagnoses) with ≥1 overnight stay in HES, identified using HES Admitted Patient Care (APC) file.
o Total days of hospitalisation: The total days will be defined as the sum of the number of days between the admission date and the discharge date for each hospital admission (including all admission diagnoses), that occur in the inpatient setting.
o Total Number of stays in intensive/critical care ward: Total number of stays with ≥1 overnight stay in critical care or intensive care unit (ICU) recorded in HES critical care*
o Total days of stay in intensive/critical care ward: The total days will be defined as the sum of the number of days between the admission date and the discharge date for each intensive/critical care ward recorded in HES critical care* with variables “epistart” and “epiend” in HES critical care*
o Total Number of elective (planned admissions): For each hospital admission, the total number of elective admissions during the follow-up period for a planned admission recorded in HES APC (variables: ‘admidate’ and ‘elecdate’)
*Note: Adult critical care is covered via an extension to HES APC, but is accessed separately. They are referred to as either Augmented Care Periods or as the Critical Care Minimum Data Set; as the underlying data standards have changed over time. The source of HES critical care data is the Critical Care Minimum Data Set (CCMDS), which includes records for critical care periods in adult designated wards. Using the HES Critical Care data, the following outcomes will be measured:
• Total number of Advanced Cardiovascular Support days
• Total number of Advanced Respiratory Support Days
• Total number of Basic Respiratory Support Days
• Total number of Dermatological Support Days
• Total number of Gastro-Intestinal Support Days
• Total number of Liver Support Days
• Total number of Neurological Support Days
• Total number of Renal Support Days
Accident & Emergency (A&E) HCRU
o Accident & Emergency visits: All visits to the emergency department:
Number of visits: variable ‘arrivaldate’ recorded in HES A&E
Disposition: variable ‘aeattenddisp’ recorded in HES A&E
Duration of stay in the emergency department: variables ‘arrivaldate’ and ‘depdur’
Outpatient (speciality) HCRU
o Outpatient visits: variable ‘apptdate’ and ‘attended’ recorded in HES outpatient
o Outpatient procedures: variables ‘opertn_0X’ record in HES outpatient
o Outpatient specialty: variable ‘mainspef’ and ‘tretspef’
Luis Vaz - Chief Investigator - Bristol-Myers Squibb Pharmaceuticals Limited - UK ( BMS )
Harmony Omeife - Corresponding Applicant - Evidera, Inc
Achim Wolf - Collaborator - Evidera Ltd - UK
Audrey Jung - Collaborator - Evidera, Inc
Hanna Kew - Collaborator - Bristol-Myers Squibb Pharmaceuticals Limited - UK ( BMS )
Nahila Justo - Collaborator - Evidera Ltd - UK
Nicola Sawalhi-Leckenby - Collaborator - Evidera, Inc
Robert Donaldson - Collaborator - Evidera Ltd - UK
Shital Vekeria - Collaborator - Bristol Myers Squibb - Europe ( BMS )
Alison Isherwood - Collaborator - Evidera, Inc
Johann Castaneda - Collaborator - Evidera Ltd - UK
Luis Vaz - Collaborator - Bristol-Myers Squibb Pharmaceuticals Limited - UK ( BMS )
HES Accident and Emergency;HES Admitted Patient Care;HES Outpatient;NCRAS Cancer Registration Data;NCRAS Systemic Anti-Cancer Treatment (SACT) data;ONS Death Registration Data;Patient Level Index of Multiple Deprivation