Introduction

It is a fundamental aspect of a surgeon’s practice that he or she has an understanding about his or her own individual outcomes. This has become a key element in the recertification of surgical practitioners. Outcome measurement usually begins with some form of self-audit. However, the ability to benchmark one’s own performance against peers is a more powerful measure of surgical performance. But who does one benchmark against? One option is published outcomes in surgical journals, but there is an inherent bias in these publications, as no one strives to publish poor results. An alternative approach is a national registry, and advances in technology have enabled a greater ease of data entry and data analysis. In addition to evaluating clinical outcomes, such registries can map trends in clinical activity and review clinical practice. This chapter explores the background of registries, explains how to establish a national registry, and details the outcomes from the British and Scandinavian experience.

Measurement of Quality: The Science of Improvement

In 1990, the Institute of Medicine defined quality as “the degree to which Health Services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge.”¹ Most doctors regard providing high-quality care as not only a professional responsibility but also as their raison d’être. For surgeons, the focus is on patient/populations and clinical effectiveness and safety. However, the wider dimensions of quality such as efficiency, equity, patient-responsiveness, access and coordination must not be overlooked.² Audit is a major component of the science of improvement, and a reliable system of measuring outcomes has many benefits:

Surgical Registries

A registry is defined as a systematic collection of a clearly defined set of health and demographic data for patients with specific health characteristics, held in a central database for a predefined purpose.³ Compared to clinical trials, registries try to include the whole population with a certain disease or operation without any selection criteria. The disadvantage of this form of research is that the amount of data collected is small compared with clinical trials. However, registries aim to obtain a complete and unbiased overview of all cases. There are two main outcomes of a registry:

• A tool that monitors and improves quality of care

• A resource for epidemiologic research

There are excellent examples of local endocrine databases which have produced data that have brought about improvements in the surgical management of endocrine disease.^4,⁵ One of the best examples of long-term data acquisition that covers an entire nation is the Swedish Cancer Registries, which has made long-term population-based research possible, for example, on prognostic factors and survival rates in differentiated thyroid cancer.⁶ However, there are fewer examples of registries that include surgical treatment, the exception being the Society for Cardiothoracic Surgeons of Great Britain and Ireland.⁷ The principal advantage of running such an initiative at a national level is that it enables individual surgeons to benchmark their own practice with their peers.

Steps in the Development of a Registry

There are two procedural divisions: the central coordinating unit and the local center. The central unit sets up and maintains the registry, whereas the role of the participating local centers is to enter the data. One key driver for success is a leader who champions the project. The intended use of the registry data determines the necessary properties of the data, so it is fundamental that the designers of a registry have a clear preconceived understanding of the intention of the registry. So at the outset it is absolutely fundamental that clear objectives are written. We advise that any new venture should commence with a pilot project.

The most important step is defining a sensible minimum data set (Table 48-1). It cannot be emphasized enough that there is a balance to be achieved between collecting sufficient data and avoiding the creation of an unwieldy system that takes too long to enter the data and risks losing compliance from the local centers. The data can be used for multiple purposes (Figure 48-1). In fields of surgery where mortality and major morbidity are a significant feature, it is important to establish a risk-adjusted benchmark for clinical outcomes, which addresses the patient’s health status or functional ability such as the Physiological and Operative Severity Score for the enUmeration of Mortality and morbidity (POSSUM) and National Surgical Quality Improvement Program (NSQIP).⁸ For thyroid and parathyroid surgery, this detail is probably unnecessary so long as variables including age, reoperative surgery, pathology, and retrosternal extension (for thyroid) are recorded. Another key step is to choose a software company that can provide appropriate technology to minimize errors and maximize the efficiency of obtaining results. Nowadays this includes Internet capability as well as data collection and interrogation. There are two types of data errors: systematic and random.⁹ Systematic errors are caused by programming errors, unclear definitions for data items, or violation of the data collection protocol. The software should therefore be designed to be user-friendly and should include definitions whenever possible. Drop-down boxes and “hover” buttons are particularly useful in this regard. Random errors are caused by inaccurate data transcription and typing errors. These can be minimized by using ambiguity-free text.

Table 48-1 Composing a Minimum Data Set

Data Set	Examples (Thyroid)
Demographics	Age, gender
Risk factors	Reoperative surgery, retrosternal goiter
Clinical history	Endocrine status, compressive symptoms, hereditary disease
Pre-op investigations	Vocal cord check, cytology
Operation details	Principal surgeon, nodal surgery, use of nerve monitor, extent of surgery
Pathology	Primary pathology, incidental pathology
Early complications	Hypocalcemia, hemorrhage
Long-term outcomes	Voice change, hypocalcemia