The Demography Myth: How Demographic Forecasting Underestimates Hospital Admissions, and Creates the Illusion that Fewer Hospital Beds and Community-based bed Equivalents, will be Required in the Future

Abstract

Aims: This Policy Article, which includes data synthesis, seeks to evaluate the effectiveness of demographic-based forecasting of future hospital admissions, and hence for hospital bed numbers. A role for the absolute number of deaths, as a proxy for persons approaching the end-of-life, will also be investigated, especially as this is the principle factor behind the volatility in bed occupancy. Study Design: Literature review plus supporting analysis of relevant trends. Place and Duration of Study: Studies from a variety of countries, analysis of data relating to the NHS in England, additional analysis relating to the Kings College Hospital, London. Findings: Demographic forecasting is subject to the constant risk fallacy, namely, admission rates are changing over time. A variety of factors can be seen to lead to changes in admission rates. There is strong evidence that it is the trend in the absolute number of deaths, rather than demography which drives most of the marginal changes in bed demand. Up to 55% of a person’s lifetime bed utilization may occur in the last year of life. However, the ratio of deaths per 1,000 population ranges from 4 to 16 (inner city communities to more rural and retirement locations). Deaths per GP also ranges between 5 to 27, while the crude mortality rate in English hospitals ranges from 1.3% to 5.4%. These imply that in some locations end-of-life has a greater impact not only on primary, secondary and social care resources, but also on the volatility in resource requirements. There are 20 deaths per GP in the Torbay integrated health and social care organization indicating that the 30% fewer bed days per death achieved in this organization may not be replicated elsewhere. Same-day-stay admissions require separate forecasts since they rely on technological trends driving day surgery rates (elective admissions) and the need for observation and rapid diagnosis in medicine (both elective but mainly emergency admissions). In England, an obsession with fewer beds on behalf of the Department of Health (now NHS England) has been a key factor bringing the NHS into crisis. The worst affected hospitals now run at close to 100% average occupancy. However, the period 22 weeks prior to death appears to mark a watershed for acute care, where surgical, medical and critical care interventions become increasingly futile. A shift to a more palliative-based model of non-acute care is recommended, and will lead to considerable savings in occupied beds. A new generation of models are desperately required, as are supporting biochemistry, vital sign and frailty-based algorithms to enable care to be diverted out of acute hospitals during the last weeks of life.