Capturing the configuration of hospital-based services: a comparison of national surveys and individual patient data for liver disease in England

National surveys of health care providers are increasingly being used to assess if complex disease service configurations meet the needs of specific populations. We evaluated one-off national surveys of hospitals in the English National Health Service and compared their results against routinely collected individual patient data, using liver disease services as an example. We identified one-off national surveys describing liver disease services provided by NHS hospitals through stakeholder communication and developed a content framework, and compared information about patient flow, capacity, workforce, and specialist service provision against a number of routinely collected datasets. Nine national datasets were identified that provide information about liver disease services with notable levels of duplication. Four datasets were derived from one-off hospital surveys. Five were routinely collected, and two of these included individual patient data. Whilst providing some unique detail, the one-off surveys had inconsistent service definitions, inferior response and completeness rates, and marked discrepancies, compared to routinely collected data. Response and completeness rates were highest for the one-off survey with the lowest number of items and quantitative metrics. Response and completeness rates of the routinely collected datasets were always near-100%. When assessing the national configuration of complex disease services, one should first consider if routinely collected datasets already provide the required information. If used, one-off hospital surveys should avoid unnecessary collection of data, ensure consistent service definitions used in other relevant datasets, and optimise response and completion rates.