Melanomas Excised in Primary Care vs Secondary Care Excision

Cutaneous melanoma is increasing in incidence. UK incidence has quadrupled since 1970. With corresponding increased public skin cancer awareness, GPs are frequently asked to examine pigmented lesions. Definitively diagnosing melanoma on clinical features alone is extremely challenging and many lesions exhibiting worrying features require a skin biopsy. Existing guidelines discourage initial primary care biopsy of suspicious skin lesions despite evidence that it leads to earlier diagnosis and may benefit patients and the NHS.

Results from Previous Research To the investigators conducted the first blinded study comparing relative performance of primary and secondary care in excising melanomas and found patients receiving initial excision of melanoma in primary care were no more likely to have it inadequately excised than those initially biopsied in secondary care. In a further study, we found that patients receiving their initial diagnostic excision biopsy for melanoma in primary versus secondary care were no more likely to be dead, or to have died of metastatic malignant melanoma. Patients who had their initial diagnostic excision biopsy for melanoma in primary care had significantly fewer subsequent hospital admissions and spent fewer days in hospital.

Aims

1. To determine whether there is any difference in mortality for people diagnosed with cutaneous melanoma in Scotland between 2005-2013 following initial excision in primary versus secondary care.
2. To determine whether there is any difference in morbidity following initial excision in primary versus secondary care.

Research Questions

1. Is there a difference in survival between people having initial excision of melanoma in primary versus secondary care?
2. Is there a difference in subsequent risk of death from recurrent melanoma between those having initial excision of melanoma in primary versus secondary care?
3. Do people having initial excision of cutaneous melanoma in primary care versus secondary care subsequently have more hospital admissions?
4. Do people having initial excision of cutaneous melanoma in primary care versus secondary care subsequently have more outpatient attendances?

Methods Data from four national datasets will be linked: the Scottish Cancer Registry (SMR06); general/acute inpatient and day cases (SMR01), outpatients (SMR00) and the GRO(S) death registry. Following approval via the PBPP mechanism the data will be linked and provided to the research team by Dr Fiona Campbell and team and the Information Statistics Division of the NHS in Scotland, Edinburgh.

The Scottish Cancer Registry is maintained by the Information Services Division (ISD) and has been collecting information on cancer since 1958. Approximately 47,000 registrations are made annually in Scotland and the database currently holds over 1,500,000. The registry has high levels of completeness and accuracy, with notifications from hospital, pathology, and death records. Date of diagnosis, 'stage' at diagnosis, (Breslow thickness and Clark levels), histological verification, tumour grade, initial treatment received and treatment dates are all recorded.

The Scottish Morbidity Record SMR01 database is an episode-based record containing all general/acute inpatient and day cases discharged from Scottish hospitals. The quality of the database has been assured. The Scottish Morbidity Record SMR00 database records all new outpatient attendances, although recording of re-attendances is variable. Diagnosis is not recorded on SMR00, but hospital speciality and outpatient procedures are. General Register Office for Scotland (GROS) Death Registry data include date of death, primary and secondary causes.

Data Linkage Following appropriate approvals all individuals in the Scottish Cancer Registry recorded with cutaneous melanoma between 2005- 2013 will be linked to SMR01, SMR00, and GRO(S) using the CHI number, a unique patient identifying number issued to every person registered with a GP in Scotland used to link health data for research purposes.

For all individuals with cutaneous melanoma, SMR01 data for 10 years prior to cancer diagnosis will be used to construct co-morbidity scores in two ways. First, principal and supplementary diagnostic codes from hospital attendances and admissions will be used to construct Charlson scores. These scores are used to predict prognosis based on age and comorbid conditions. With each increased level of the comorbidity index the cumulative mortality attributable to comorbid disease increases in a step-wise fashion. The Charlson score, an estimate of 10-year survival can be used in longitudinal studies. A second proxy co-morbidity score will be constructed from total number of inpatient bed days during a pre-defined interval preceding presentation with cancer.

The date of cutaneous melanoma diagnosis will be the date of issue of the pathology report, which will be available for all individuals included in the cohort.

Data Storage Once linked these data will be placed into the National Data Safe Haven where it will be accessed remotely by the members of the research team approved to access it. These data will be accessible by the team for 3 months and subsequently archived for 5 years in full compliance with the Research and Information Governance Procedures of the NHS in Scotland, ISD and the National Safe Haven.

Main Outcomes

1. All all-cause survival from the date of melanoma diagnosis.
2. Cause-specific survival (including recurrent melanoma) from date of melanoma diagnosis.
3. Total number of hospital attendances (inpatient and outpatient visits) and total number of days from date of melanoma diagnosis.

Main Exposure Location of primary excision, (primary or secondary care) - identified from electronic histopathology records.

Other Independent Variables Patient level: age, gender, residence category, deprivation , Charlson co-morbidity score; inpatient bed days; type of biopsy type; anatomical site; melanoma type; Breslow depth and Clark level.

Operator level: operator specialty (plastic surgeon, dermatologist, other hospital specialist, GP frequent exciser, volume of biopsies submitted.

Statistical Analysis Analyses will be conducted using a hierarchical, multilevel framework, with patients nested within operators. Estimated survival probability at the end of follow up will be obtained. Univariate Cox regression will be used to compare survival probabilities between individuals receiving initial diagnostic excision biopsy in primary versus secondary care. Cox proportional hazards regression, with robust standard error estimates, will be used to explore associations between patient-and operator-level factors with all-cause and melanoma related mortality.

To explore morbidity, number and duration of admissions and outpatient attendances (all and individual specialty) will be calculated for each individual for each year following diagnosis. Association between location of initial diagnostic excision biopsy and morbidity will be examined using GLMs with (a) a Poisson distribution and log link function for outcome factors which are counts (b) a Binomial distribution and log link function for outcome factors which are binary, and (c) a Gaussian distribution and identity link function for outcome factors which are continuous. Models will be unadjusted and then adjusted for patient-and operator-level factors.

Analyses will use Stata (version 13) or MLwiN software. A p-value of < 0.05 will be considered statistically significant throughout.