Introduction

INTRODUCTION

This is the third publication of the outcomes of head and neck cancer surgical treatment in English Hospitals. It reports on the patients who were diagnosed with cancer after 1st November 2012 and before 31st October 2014. As before, these outcomes are drawn from the data collected for the Data Audit for Head and Neck Oncology (DAHNO).

Head and neck cancer

The term cancer refers to any malignant tumour. The term malignant usually means that the tumour can form secondaries elsewhere in the body away from the original, or primary, tumour site. Although the term benign tumour means that it does not form secondaries away from the primary site some benign tumours can become malignant, or their growth at the primary site, such as in the brain, may cause the patient’s death. In addition there are some tumour types whose behaviour ranges from benign to malignant depending on their histological grade. DAHNO only included data on tumours that were malignant.

Head and neck cancer affects the most socially important part of our bodies. We speak, we eat and drink, and we observe and listen to others through our throat, mouth and face respectively. Furthermore, we use our facial muscles to express ourselves and our facial appearance is a representation of our identity. Head and neck cancer or its treatment may adversely affect all these aspects of our lives and even destroy our sense of ourselves.

The words head and neck are a bit misleading because brain, eye, skin and thyroid cancer are dealt with separately and not included in the head and neck cancer group.

Even so the head and neck cancer grouping is diverse enough and very heterogeneous unlike other cancer sites such as breast or colon. The many subsites of head and neck cancer are very different anatomically, physiologically and functionally. This variety, and the social aspects of head and neck cancer mean that many clinical groups contribute to the overall management of the head and neck cancer patient.

Most head and neck cancers arise from the mucosal surface of the upper aerodigestive tract, the site of the body where we take in both air and food. These are mainly squamous cell carcinomas. The sub sites of head and neck cancer are grouped into those affecting:

• The lips;
• The front of the mouth (oral cavity) which itself includes the upper and lower jawbones, the inside of the cheeks (buccal), the mouth floor (stretching between the tongue and the gum fixed to the inner aspect of the lower jaw behind the teeth), the gums overlying the upper and lower jaws, the front visible part of the tongue (anterior 2/3rds) and the hard palate which is bony hard, fixed and immobile.
• The inside of the nose
• The sinuses. These are the air sacs in the hollowed out bones adjacent to and above the nasal passages (the maxilla, ethmoids and frontal sinuses). In normal life these sinuses help warm and moisten the air we breathe.
• The back of the mouth (oropharynx). This includes the tonsils, the back part of the tongue that is not visible (the tongue base) and the soft movable part of the palate
• The whole pharynx which is essentially a tube which is incomplete at the front. It starts at the top as the nasopharynx lying behind the nasal cavities in front and the bony base of the skull behind. Here air passes from the nose into the pharynx. The pharynx continues down to the oropharynx which is open at the front to the mouth. The pharynx then extends down and splits in two as the laryngopharynx and cricopharynx leading to the trachea (windpipe) and oesophagus (gullet) respectively. The hypopharynx connects the cricopharynx above to the oesophagus below
• The glands that produce saliva. There are multiple very small salivary glands throughout the mouth and throat and 3 pairs of large glands on each side. One pair under the skin in front of the ear (parotid glands); one pair under the lower jaw at the back (submandibular glands); and one pair under the tongue (sublingual glands). The cancers that occur here are mainly adenocarcinomas.

The first place that head and neck squamous cell carcinomas spread to (secondaries also called metastases) is the neck glands.

There are other less frequent types of malignancy affecting the head and neck such as mucosal melanoma and sarcoma. Sarcomas are rare and are treated in a very few centres where all body sites of sarcoma are treated. They have never been reported in DAHNO.

The main treatment modalities for head and neck cancer are surgery, radiotherapy and chemotherapy. New immunological agents and antibodies are being developed and introduced as adjunctive treatments. Approximately 2/3rds of patients will be treated with radiotherapy or chemoradiotherapy either as primary treatment or as adjunctive treatment after surgery. The audit has always only concentrated on surgical treatment.

Historically, DAHNO focused on surgical treatment used as the first option to treat the patient and the immediate results of this treatment. The two previous Consultant Outcomes Publications (COP) only evaluated surgery when it was the first treatment and did not evaluate surgical treatment after prior radiotherapy or chemotherapy, or surgery for recurrent disease. It only reported on malignant cancers.

History of DAHNO

The DAHNO national registry was founded over 14 years ago by the British Association of Head and Neck Oncology (BAHNO) (website details), led at that time by Prof John Watkinson, an ENT surgeon in Birmingham. BAHNO set out to create an audit or register of the practice of head and neck surgery around the UK in terms of the number of hospitals and surgeons who were treating head and neck cancer, and how they investigated and treated their patients. BAHNO then used this information to set standards and guidelines (reference).

The wider National Health Service (NHS), which had been documenting cancer incidence and prevalence through its regional cancer registries for over five decades, then set up and funded national audits. Therefore, in 2004 the DAHNO audit, funded initially by The Health Care Commission of the NHS, commenced in earnest along with one other national cancer audit examining Lung Cancer. It was restricted to data collected from England and Wales.

Over the next 10 years DAHNO was run by the NHS Information Centre (then called the Health and Social Care Information Centre – HSCIC - till this year, when the name was changed to NHS Digital) with clinical leadership by 2 surgeons. Slowly but surely over the next 10 years DAHNO gained increasing buy-in and commitment from hospitals and NHS trusts who provided administrative support for the multidisciplinary teams (MDTs) in the form of MDT coordinators. Therefore, after a rather low take-up, estimated at one quarter of the incident cases in the first annual report, the involvement of Hospitals and Trusts improved significantly so that the 10th annual DAHNO report estimated 92% of patients captured.

Although the system of collecting data for DAHNO at weekly Multi-disciplinary meetings - MDTs (when all the interested professional groups managing head and neck cancer should be present) seemed logical, unfortunately, with a few notable exceptions, it relied on the least clinically experienced member of the MDT – the MDT coordinator – to collect and record the data at that time. Furthermore, in order to complete all the fields of the dataset on every patient, these MDT coordinators spend the whole week after the weekly meeting searching through their hospital’s computer networks for pathology and radiology reports, clinic letters and discharge summaries, then interpreting these documents to complete fields on for example, cancer staging, clinical treatment and treatment outcomes.

This system of one single person being responsible for completing the data fields made some sense as the HSCIC only needed to liaise with one person at each hospital. Unfortunately, very few if any of the clinical groups actively participated in recording data relevant to their activity believing that this responsibility lay totally with the MDT coordinator. Therefore it is not surprising that although pretty well every head and neck cancer patient is now entered into DAHNO most of these records are incomplete; for example, less than half of the operation records had a record of which surgeon performed them.

The Consultant Outcomes Publications 1st November 2011 to 31st October 2013 (COP 1 and 2)

The HSCIC and DAHNO’s clinical leaders collected the overall raw DAHNO data from each hospital / NHS Trust once yearly after 1st November.

Linkage with other data collections held by the National Cancer Registration Service - NCRS (now called the National Cancer Registration and Analysis Service - NCRAS) and HSCIC itself was used to check and augment the DAHNO data. For instance, these data collections were used to check that all new cases of head and neck cancer for the treatment period had been captured by DAHNO (case ascertainment) and that a consultant surgeon’s name was attached to each surgical operation.

In addition, the data was sent back to the respective hospitals cancer managers and MDT leads / coordinators. The surgeons and their MDT Leads were given a period of 7 weeks in the first COP, then 12 weeks in the 2nd COP to correct mistakes and augment data fields that were empty (e.g. surgeon/operation fields).

Appendix 1 of the 2nd COP report included a clear statement for each NHS Trust showing the number of operations that could not be linked to a surgeon. Some individual surgeons performed heroic efforts over the 12-week window to ensure that every operation could be linked to a surgeon in their Trust. This occurred in 21 out of 94 Trusts. Conversely almost 60% of operations could not be linked to a surgeon in 18 out of the 94 Trusts.

In these first and second COP reports the authors issued several words of caution regarding interpretation of the data:

• Data completeness issues should be factored into any reading of the analysis.
• Please note that this list is not intended to be a definitive list of head and neck cancer surgical hospitals.
• As the Audit was not originally designed to report at surgeon level, a new system was developed to allow individual surgeons to assign activity to themselves
• Surgeons, who were given seven weeks to update data onto the system, supplemented information for just under half of these procedures (3409 of 7195, or 47 per cent).
• Some organisations had difficulties or were unable to supplement data about their activity. This should be taken into account when interpreting the data.
• Due to the relatively low number of these specific procedures undertaken by each surgeon an increase in one or two deaths can have a large impact on the surgeon’s mortality rate.
• It is essential to maintain a balance between transparency and patient confidentiality in respect of publishing information. This is especially important when releasing data where there are very small numbers. In order to protect patient confidentiality, we have put controls in place to limit the level of information disclosed. For some surgeons this greatly limits the utility of the data.

COP 3 - The 2012/2014 Clinical Outcomes Publication

In July 2015, the Health Quality Improvement Partnership (HQIP) (www.hqip.org.uk) on behalf of NHS England and the Welsh Government awarded the prospective national head and neck audit contract to the surgical research charity, Saving Faces–The Facial Surgery Research Foundation (www.savingfaces.co.uk), who were working in collaboration with BAHNO. The technology partner for this new head and neck audit (HANA) is Dendrite Clinical Systems (www.e-dendrite.com) who run many other national and international audits funded by the individual national specialty organisations. Two particularly relevant audits run by Dendrite are the UK National Flap Registry (uknfr.e-dendrite.com), recording the outcomes of reconstructive surgery in the United Kingdom, and the Thyroid and Parathyroid Surgery Registry (https://cl2.n3-dendrite.com/csp/baes/frontpages/index.html) on behalf of the British Association of Endocrine and Thyroid Surgeons (BAETS). This consortium continued the work performed by the HSCIC whose contract finished on 31 October 2014.

Although analysing historic DAHNO data for the period before 1st November 2014 was not part of the original tender and bid, HQIP asked Dendrite to take on the generation of the 2013/14 COP as HSCIC had declined to produce this report on the data they had collected and held.

All the historic DAHNO data covering the 2013/14 data period was still held by HSCIC so it was necessary for this data to be transferred to Dendrite before any analysis could take place. But first Dendrite had to obtain permission to use patient identifiable data without consent and therefore demonstrate to the Confidentiality Advisory Group (CAG) of the NHS research arm that satisfactory information governance systems were in place before the NHS would agree to Dendrite receiving this identifiable patient data from HSCIC. We gained this approval on 15th September 2015.

This turned out to be only the first step in the data transfer. We then had to negotiate the HSCIC Data Access Request Service (DARS) and seek approval from HSCIC before they would transfer any data. Without this transfer no COP analysis could be conducted by the HANA team.

We did not obtain any DAHNO data from HSCIC for six months till January 2016. Even then, the data transferred from HSCIC had been stripped of any linked Office of National Statistics (ONS) mortality or Hospital Episode Statistics (HES) data.

Our original application was reviewed several times by HSCIC who implement a multi-stage checking process even before their Data Access advisory Group (DAAG) review of the application. At each stage we needed to revise our application and DAAG have now requested that Saving Faces demonstrates that Saving Faces has appropriate system level security.

This has resulted in 2 alterations to the reported measures in COP 3 compared with the previous two COPs. Less than 40% of the records received from HSCIC had a surgeon name associated with an operation. We are therefore reporting at hospital rather than surgeon level. Without ONS linkage we have not been able to report on patient mortality within 30 days of surgery, but we have been able to report on whether a patient died in hospital after a surgical operation.

COP 2 reported on 2 years of DAHNO data. One of those years is included in both COP 2 and our COP 3 – 1st November 2012 to 31st October 2013. After filtering the DAHNO data received from HSCIC to exclude ambiguous and unreliable data we identified 5,594 surgical records for analysis. However, this number was only around 80% of the number of cases (6,830) that were included in the 2nd COP report. This confirms that our COP 3 report is likely to be less comprehensive than the COP 1 and 2 reports and must be interpreted with even more caution than the authors of the COP 1 and 2 recommended.

As was mentioned in the 2nd COP report surgeons and hospitals were given 12 weeks to improve on the raw DAHNO data for their hospital. This COP 2 report was delivered in November 2014, twelve months after the original DAHNO data had been collected. Although we have only been working on the data we received from HSCIC for 7 months, COP 3 is already 8 months behind in the reporting cycle. Therefore, we cannot offer this period of 12 weeks but have settled for a lesser period of time to allow hospitals and surgeons to review and adjust their data if they wish to improve its accuracy. We recognize this is doubly challenging as several surgeons will have left the hospital or stopped performing cancer surgery, new surgeons for whom the data is not relevant will be in post and some of the data is on patients who were treated almost 4 years ago.

In summary this COP 3 report has been conducted on data between 2 and 4 years old from a database not designed by HANA, with data not collected by HANA at source, and delivered by the previous contractor without valuable linkage data. Several of the current surgeons working in the hospitals were not there when this data was collected. However, although the data for COP 3 is likely to be incomplete we believe that the 8 fields reported on provide enough valuable information for patients, hospitals, clinicians, commissioners and the NHS to make publication worthwhile, although the data must be interpreted with care. We will report on 5 fields for NHS Choices and MyNHS and 8 fields will be displayed on the COP 3 2016 Web Portal.