Boom or bust? Estimating blood demand and supply as the baby boomers age

On January 1, 2011, the first-born of the baby boomer generation celebrated their 65th birthday. The social and economic impact caused by the aging and retirement of those born between 1946 and 1964 has long been debated and is now upon us. For the blood banking industry in the United States, little has been discussed publicly with even less preparation for the coming changes. US blood centers appear oblivious to the possible effect of shifting demographics and are poorly prepared to forecast long-term blood demands, raising the specter of future blood shortages even as the industry wallows presently in an apparent glut of blood. While the National Blood Collection and Utilization Survey (NBCUS)1 has proven a powerful tool to track blood donation, and the US Census Bureau has mapped and predicted the effects of demographic changes well into the future,2,3 inadequate information on the demographics of blood use in our hospitals and those factors that influence utilization has left us without the tools we need to inform policy. In this setting, the report of Greinacher and coworkers4 sounds a warning bell. In the small German federal state of Mecklenburg-Pomerania with a population of 1,707,266 inhabitants, the authors were able to track the vast majority of hospital transfusions and blood bank donation records for a 1-year period (2005). These were used to integrate age- and gender-specific transfusion rates with well-established population demographics to predict the influence of population changes on the relative blood supply. The process was validated by utilizing 1997 transfusion data from one hospital to predict blood demand and supply in 2007. Remarkable accuracy was demonstrated, with a difference of only 226 red blood cell (RBC) units (1.2%) from that actually used. In this former East German state, marked population changes followed the fall of the Berlin Wall and are predicted to continue into the future. Birth rates have fallen precipitously and young people often migrate elsewhere to follow financial opportunities, while the postwar generation ages and is expected to live longer than their parents. Overall the population is predicted to decline by 8.3% by 2020, but disproportionately: the potential blood donor group aged 18 to 68 years may decline by 16.1% while the patient group 65 years and older that received approximately 62.1% of RBC transfusions in 2005 is predicted to grow by 26.4%. Assuming no changes in donation patterns and medical indications for blood use, these opposing forces are predicted to yield a 47% shortfall in the blood supply by 2020, representing a major challenge to blood providers and their hospital users. This study has weaknesses: Age-specific blood use was estimated using data from major hospitals that represent only 80.5% of distributed products. It is possible, but unlikely, that transfusions occurring outside the hospital environment would show differing demographics. It is also based on only a single year of data, which may be subject to peculiarities. Finally, a study of this nature cannot account for new technologies, blood management strategies, or world events (e.g., global recession) that may affect supply and demand. Nevertheless, it is clear that population demographics provide a powerful baseline on which to build future plans for the blood supply and that the need for such planning is critical. Other countries show similar population dynamics to those described in Germany. Indeed a number of major economies, including Japan, Hong Kong, South Korea, Italy, Spain, and many other European countries have inverted population pyramids that bode poorly for the future blood supply, with dramatically increasing aged groups and diminishing young donor pools.2 Interventions to increase the blood supply and decrease blood utilization are needed sooner rather than later. How do these data reflect the US situation? Population demographic predictions are readily available, calculated from the 2000 census that will soon be updated with 2010 data.2,3 The threat is real: the US population is expected to swell by approximately 10% over the next 10 years; however, the major donor population (aged 16-64) will increase by only 5.2%, while those 65 years and older will increase 36.2% from 40.2 million to 54.8 million. The 65 years and older age group will shift from 13.0 to 16.1% of the total population, an absolute 3.1% increase, dramatic indeed if this group uses the majority of the blood supply (Fig. 1). Unfortunately, little US age-specific blood use data are available. Early work performed in a limited setting in Olmsted County, Minnesota, in 1989 through 1992, suggested that 53.3% of RBC units were transfused into patients older than 65 years at a time when the incident rates of RBC transfusion were approximately 48.7 units/1000 inhabitants.5 The findings of Greinacher and colleagues that 62.1% of RBC were transfused into this age group is consistent with similar reports in other countries. In Finland, the average RBC use at age 55 is close to the national average of approximately 50 units/1000, but increases in a sigmoid fashion to more than 200 units/1000 population by age 90 years.6 Similarly, in the northern part of England, where there was a relatively low per-capita blood use of 42.7 RBC units/1000 in 2000, the average RBC transfusion rate above the age of 65 years was approximately 151 units/1000, but only 21.8 units/1000 for patients aged less than 65 years. A total of 57.2% of RBCs were used by patients 65 years and older.7 Similar dynamics are evident in small studies performed in Western Australia and in national database studies in Denmark and Sweden.8,9 Although precise data are needed, all indications are that in the United States, the majority of RBC units are transfused to patients 65 years and older and that this may constitute 55 to 60% of the blood transfused. Given a predicted 36.2% increase in patients 65 years and older and a 6.1% increase in patients less than 65 years old, absolute RBC needs may increase by approximately 22.8 to 24.3% over the next 10-year period, or an average of 2.3 to 2.4% per year based on demographics alone (assuming no changes in medical practice or other drivers; Fig. 1). Using a rough estimate of future collections based on the number of collections in 2008 (from the 2009 NBCUS)10 and the rate of growth of the major donor population aged 16 to 64 years (5.2%), US transfusion demand will exceed collections by 2020 (Fig. 1). These projections do not take into account reported declines in blood usage since 2008 related to the economic recession and/or implementation of blood management strategies. Neither do they account for the increasingly ethnic heterogeneity in the US, where minority populations that have not supported blood donation enthusiastically in the past will comprise a larger proportion of the potential donor pool. Changes in the age-specific US population based on the 2000 census 2 and projected RBC collections and transfusions extrapolated from 2008 collection and transfusion data 10 (assuming that 55% of RBCs are utilized by patients ≥65 years old). (▪ ) Less than 16 years old; (◆) 16 to 64 years old; () 65 years and older; (▵) units collected (projected); (×) units transfused (projected). The question whether the US blood supply can sustain a substantial period of growth is particularly difficult to contemplate, given the recent recession and reports of decreased demand and a blood "glut." Demographic changes have not previously affected the US blood supply on a major scale and represent a novel, underappreciated dynamic; nevertheless, history demonstrates that the United States has grown its blood supply on this scale before. After declining blood utilization during the post-AIDS period, the NBCUS reported a 33.9% increase in transfusions and a 42.4% increase in whole blood collections between 1997 and 2008.1,10 In those 12 years the overall population increased by 11.6% while the 65 years and older populations increased by 12.8%. Growth in blood use was driven predominantly by expansion of medical and surgical use rather than by population demographics, highlighting the difficulty we now face: Changes in medical practice can affect blood use on the same scale as population demographics, making predictions difficult in the short term. Analysis of how the United States managed the 1997 to 2008 expansion is informative: During this period increased blood donation was driven by more efficient use of current donors rather than by widespread expansion of the donor pool. The licensing of 2-unit RBC (2RBC) collection technologies in 1997 led to the rapid increase in double units collected, comprising as much as 11.3% of the supply in 2008. In the American Red Cross system, Zou and colleagues11 demonstrated that fewer donors donated whole blood in 2005 than in 1995, but those donors generated 11.8% more units, presumably by increasing the donor return rate and by the use of 2RBC collections. Nationally, donations from high schools and colleges expanded to provide 19% of the total blood supply in 2008. Donors aged 16 to 24 years donated at a rate of 84.8 units per thousand donors,10 as more and more states allow 16-year-olds to donate with parental consent. These young donors hold the key to increasing the depth of the donor pool in the years to come, as it is unlikely that future growth of collections can be sustained without substantial increases in the numbers of active donors. Blood management programs and changes in medical practice, demand for elective surgeries, and surgical hemostatic technique may all impact blood demand to magnify or diminish the effect of population changes. The United States, however, is poorly equipped to detect these changes, due to the lack of data on blood use at the patient level. This dearth of knowledge is highlighted by the report of Rogers and colleagues,12 who provide a snapshot of blood use in the older US population by integrating data from the University of Michigan Health and Retirement Study (HRS) with Medicare files. Medicare data, although the most comprehensive US database of medical care of the elderly, are limited in that they do not necessarily represent the overall population and, more importantly, transfusion records may be incomplete due to complex billing practices under the diagnostic-related group system. The HRS study of individuals over 50 years old provides a statistically sound, nationally representative cohort that has been interviewed every 2 years from 1991 to 2007. A group of 16,377 patients older than 65 years from 1992 through 2006 were included in this report, with survey weighting to represent the US fee-for-service Medicare population. Thirty-one percent (95% confidence interval [CI], 30%-33%) of Americans older than 65 years received at least one transfusion within a 10-year period and 5.8% (95% CI, 5.4%-6.2%) experienced repeated transfusion-related visits to hospitals or health care providers within 30 days. The average number of transfusion-related visits was 2.3 over a 10-year period (95% CI, 2.2-2.4). When predictors were considered individually, adults 85 years of age or older, men, African Americans, Mexican Americans, underweight adults, smokers, nondrinkers, and those without a high school degree were more likely to have received blood. Transfusion was more frequent in Southern regions (34% of older Americans receiving transfusions in a 10-year period) and the Midwest (33%), intermediate in the Northeast (30%), and lowest in the West (26%). An older person living in the South was 3.14 times more likely to have received a transfusion during a 10-year period than one living in the West, independent of age, sex, race, various personal characteristics, and health-related conditions, suggesting that transfusion practices of providers in one region may differ from practices in other regions. These findings highlight the variability in RBC transfusion demands and practices, but the lack of patient-specific demographics, including transfusion indication, age at transfusion, sex, and race precludes linkage to census data and a more accurate assessment of future blood needs. Blood collection and utilization data are available through the NBCUS, and soon the newly established Biovigilance Component of the National Healthcare Safety Network at the Centers for Disease Control and Prevention will generate aggregate data from this voluntary surveillance program. This system will specifically monitor adverse reactions and near-miss incidents related to transfusion, including denominator data for the transfusions performed in participating hospitals. Clearly, there is a need to consider more comprehensive monitoring of blood use at the patient level, either through the introduction of systematic national data collection programs or through studies using large databases from both public and private sources. To assess blood donor demographics, the voluntary donor hemovigilance program will collect demographic denominator data from blood centers in the United States, including age and sex of donors, allowing a better understanding of age-specific donation patterns that can be linked to the US Census and provide US blood collectors advance warning of gaps in supply. In summary, the United States may very well face the stark baby boomer reality of many of our industrialized colleagues, although perhaps not so immediately. We are likely to be condemned to periods of shortages and excess through our failure to understand the basic drivers of blood supply and demand. Demographic changes are a major influence on both the demand and the supply side of transfusion; predicting the effect of these changes will require more comprehensive collection of patient-specific demographics and the medical indications for transfusion. Only then will we be able to fulfill our mission of ensuring a safe and available blood supply for every patient in need. None.