Planning for what? Challenging the assumptions of health human resources planning
Introduction
Health human resource (HHR) planning has traditionally been a demographic exercise focusing on the size and demographic mix of provider and patient populations [1], [2], [3]. Typically, projections of future service requirements of populations, and hence the health human resource requirements are determined by current levels of health care use combined with standard population projections. These projected requirements are then compared with projections of the provider population to estimate future human resource shortfalls or surpluses [4], [5], [6], [7], [8]. Variants of this approach include estimating provider requirements to maintain current provider population ratios [9]. Typically these approaches recognize that the need for health services depends on the age and sex composition of the population, as well as population size. In particular, the need for health care generally increases with age, while the pattern of this increase varies by sex. Accordingly, health care utilization by age and sex groups is combined with population projections of the size and demographic structure of the population. Some approaches also recognize that practice patterns of providers vary over time and that the service requirements by age and sex may change (usually increase) with the development of new technologies [9], [10], [11], [12].
Birch et al. [3] developed an extended analytical framework to define future provider requirements which moves beyond demography. It identifies a fundamental limitation of the traditional demographic approach to HHR planning: the assumption that age and sex alone adequately summarize the level of need for services. Although generally associated with more direct determinants of need such as mortality and health status, age and sex are not direct measures of the need for health care services. They are merely indirect proxies for the health problems that ultimately determine the need for health services and the association between differences in age and needs for care may change over time.
As a result of changes in health determinants, the age pattern of morbidity and mortality varies considerably between populations and within populations over time [13]. Although the prevalence of health problems and risk of mortality increases with age, the progression of health problems with age is changing as the distribution of health status and risk of death by age reflects the life experiences of persons born at various points in time (cohorts). Cohorts experience different patterns of exposure to social conditions, infection, lifestyle risks (e.g. smoking) and access to effective health care services resulting in differences in longevity, morbidity, and disability experiences in adulthood and old age [14], [15], [16], [17], [18]. As a result, there have been steady improvements in life expectancy and changes in the distribution of age at death [19]. This has important implications for HHR planning. For example, persons in their last year of life account for 10–33% of total health care costs while the relationship between end-of-life health care costs and age is weak and may actually decrease with age [20], [21], [22]. People are, on average, healthier throughout their lifespan with the risk of dying at any age reducing over time which is increasing life expectancies. In this way, the aging population phenomena is the outcome of improved health (and hence lower levels of need for health care) within age groups. This is consistent with the compression of morbidity hypothesis [23], resulting in reductions in lifetime cumulative levels of morbidity and disability [24], [25], [26].
While it is clear that that changes in the patterns of morbidity and mortality by age will be important determinants of future service requirements, it is less clear that these changes will be important within the relatively short time horizon of most HHR forecasting models. Because of the inability to anticipate changes in the age patterns of morbidity that will alter future requirements, the margin of error of forecasts increases rapidly with time [27]. For this reason, the time horizon of forecasts is typically limited to 10 years [28] which is a much shorter time period than the changes in the age patterns generally investigated in the literature [14], [15], [17], [18], [25]. Assuming constant levels of need within age–sex groups over relatively short planning horizons can lead to surpluses or shortfalls in health human resource requirements that accumulate over time.
Breyer and Felder [29] have noted that expenditures on medical technologies are likely to have a larger impact on future health care spending than will a healthier population. However, these findings are the result of expenditures on health care being the outcome of influences of supply of and demands for care and the interactions between these two factors. The focus of this paper is to consider how the distribution of needs for care by age group have changed over a relatively short period, irrespective of what has happened to supply and demand, and the implications that has for health human resources planning. In particular we address the following research question—Is the rate of increase of health problems with age declining between age cohorts (i.e. are people ‘aging’ at a slower rate than previous population cohorts)?
Section snippets
Materials and methods
A schematic overview of the study design and analysis is illustrated using a “lexis-diagram” [30], which describes a population by age and time (Fig. 1). The diagonals (from lower left to upper right) denote individuals born in the same year. Changes in measures of mortality, morbidity and health progressing up the diagonals describe the ‘aging’ of cohorts. Comparing these diagonals gives us insight as to whether younger birth cohorts are aging more slowly. This contrasts with traditional
Results
To help give an overall sense of trends, Table 1 shows the prevalence rates across the four needs indicators by sex and age group over the five survey cycles used in this analysis. The clearest picture for both males and females is in the mortality data where death rates rise with age but fall when comparing across survey years. Recall, these estimates are from Vital Statistics records and not subject to sampling variability. Reported mobility problems, pain and poor self-reported health also
Discussion
Our results show that although the probability of mortality, mobility problems, pain and poor self-assessed health increases with age, the relationship between age and health has changed over time. For three of our four measures, those born more recently are ‘aging’ more slowly.
Controlling for age, both males and females born more recently are less likely to die than those born earlier suggesting Canadians are living longer over time. Our analysis reveals mortality rates are shifting over time
Conclusion
Even in the short time period covered in this study, the analysis shows that health care needs by age are changing over time in Canada. Planning based on constant levels of age-specific needs would therefore generate health care workforces that could not be sustained by simply maintaining the same approaches to care for those with health care needs. Given that providers have an important role in determining the demand for services we might therefore expect an expansion of the services provided
Acknowledgments
We wish to thank the Canadian Health Services Research Foundation, Nova Scotia Health Research Foundation, Ontario Ministry of Health and Long-Term Care, Health Canada (Communication Branch and Office of Nursing Policy), Saskatchewan Innovation and Science Fund, and the Capital District Health Authority for the generous financial support that made this research program possible. We would also like to thank Arden Bell and Heather Hobson at the Atlantic Research Data Centre for their assistance
References (67)
- et al.
The SHARP way to plan health care services: a description of the system and some illustrative applications in nursing human resource planning
Socio-Economic Planning Sciences
(1995) - et al.
Life expectancy and health care expenditures: a new calculation for Germany using the costs of dying
Health Policy
(2006) - et al.
Alternative projections of mortality and disability by cause 1990–2020: global burden of disease study
Lancet
(1997) Relationship between chronic painful physical condition and insomnia
Journal of Psychiatric Research
(2005)- et al.
A study of core interventions in home nursing
International Journal of Nursing Studies
(2005) Therapeutic nursing in stroke rehabilitation: a systematic review
Clinical Effectiveness in Nursing
(2003)- et al.
Mental health service use by older adults: the role of chronic pain
American Journal of Geriatric Psychiatry
(2008) - et al.
Use of complementary and alternative therapies by patients with self-reported chronic back pain: a nationwide survey in Canada
Joint Bone Spine
(2005) - et al.
Severity of chronic pain affects health care utilization after kidney transplantation
Transplantation Proceedings
(2007) Health human resource planning for the new millennium: inputs in the production of health, illness, and recovery in populations
Canadian Journal of Nursing Research
(2002)