Ideas at the margin or marginalized ideas? Nonmedical determinants of health in
Canada
Health Affairs; Mar/Apr 2002; John N Lavis; 21, 2, 107-112

Abstract:
Beginning with the release of the Lalonde report in 1974, Canadian policymakers
and researchers have been continually exposed to ideas about the nonmedical
determinants of health. At their simplest, these ideas highlight the importance
of nonmedical factors such as income, employment, and social support in
explaining the level and distribution of health in populations. Nonmedical
factors appear to explain a large part of why some people are healthy and others
are not. The litmus test for the impact of ideas on nonmedical determinants of
health is whether they have had an impact on the health of Canadians. To conduct
this test, changes in the level and distribution of health in Canada are
described, followed by an assessment of the links in a causal chain that would
have to be established to confirm that any changes in health are attributable to
action on the basis of these ideas.

BEGINNING WITH THE RELEASE of the Lalonde report in 1974, Canadian policymakers
and researchers have been
continually exposed to ideas about the nonmedical determinants of health.1 The
core substance of these ideas has remained remarkably constant over the years.
At their simplest, these ideas highlight the importance of nonmedical factors
such as income, employment, and social support in explaining the level and
distribution of health in populations. Health care explains why some people get
well after they are sick; nonmedical factors appear to explain a large part of
why some people are healthy and others are not.3

The language used to discuss ideas about nonmedical determinants of health has,
however, changed with each passing decade. Canadian policymakers and researchers
talked about "health fields" in the 1970s, "health promotion" in the 1980s, and
"population health" in the 1990s.4 So while their U.K. colleagues framed these
ideas negatively with talk of "health inequalities," and their U.S. colleagues
tried to keep the debate on neutral terms with talk of the "socioeconomic
gradient" or "society and health," Canadians have emphasized the positive side:
working in neglected health fields, promoting health, and improving (or seeking
efficiencies in the production of) population health.5

A casual observer of the Canadian scene would be impressed by the extent to
which these positively framed ideas appear to have had an impact on Canada's
policy landscape. Federal and provincial health departments have been
reorganized to create population health divisions or business lines. These
departments have jointly created the Federal, Provincial, and Territorial
Advisory Committee on Population Health, which now uses these ideas to inform
how it presents data on the health of Canadians.6 Canada's statistical agency
used these ideas to inform the development of two new longitudinal health
surveys: the National Longitudinal Study on Children and Youth, and the National
Population Health Survey. Even health-related interest groups have repositioned
themselves. For example, the Canadian Pharmacoepidemiology Forum has become the
Canadian Association for Population Therapeutics.

A casual observer also would be impressed by the apparent impact of ideas about
nonmedical determinants of health on Canada's research landscape. The country's
principal funder of health research has been transformed over the past five
years, with nonmedical determinants of health now constituting one of the four
"pillars" of the new Canadian Institutes of Health Research and the focus for
two of its peer review committees. As well, the federal government created the
Canadian Population Health Initiative to fund research and knowledge-transfer
activities in this area. New research programs have been created within existing
organizations such as the Canadian Institute for Advanced Research, and even new
research organizations have been created, such as the Saskatchewan Population
Health Evaluation and Research Unit. New data sets have been created in the
provinces (for example, Manitoba and British Columbia) to facilitate research on
nonmedical determinants of health.

A more detailed look at the Canadian scene, however, suggests a different story.
What unites the above-cited activities is how firmly rooted they are in the
health sector. And yet the main implications of nonmedical determinants of
health lie
outside the health sector, in policy sectors (such as finance, labor, and social
services) that can influence people's income, employment opportunities, and
social supports.7 For example, policymakers in finance departments seek to
influence poverty and unemployment rates. Research about the health consequences
of economic policies could influence policy making at the margin by informing
trade-offs between policy alternatives that can be expected to achieve their
economic objectives.

The litmus test for the impact of ideas on nonmedical determinants of health is
whether they have had an impact on the health of Canadians. To conduct this
test, I describe changes in the level and distribution of health in Canada over
a
twenty-five-year period and then assess the links in a causal chain that would
have to be established to confirm that any
changes in health are attributable to action on the basis of these ideas.
Working backward from impacts on health to the
ideas themselves, the links include the influence of these ideas on policy
making outside the health sector, the presence of these ideas in policy-making
environments outside the health sector, and the clarity of the messages for
policymakers.

Changes In The Health Of Canadians, 1971-1996

If poverty, unemployment, and social isolation have been found to lead to poor
health, then action on the basis of these
ideas could be expected to lead over time to an increase (or to a slowing in the
rate of decrease) in the level of health in a population or to a reduction in
the difference in health status between rich and poor, employed and unemployed,
or socially supported and social isolated, other things being equal. But therein
lies the rub: Other things are never equal. Even the best available data sets
and analytic techniques cannot account entirely for competing explanations (for
example, improvements in the availability and quality of health care and action
consistent with but not attributable to these ideas); health-related selection
into and out of poverty, unemployment, and social isolation; and latency effects
that vary by the type of action and anticipated health impact.

Russell Wilkins and his colleagues at Statistics Canada recently examined
changes in the level and distribution of health status in urban Canada from 1971
to 1996.8 They used a variety of mortality measures (infant and adult mortality,
life expectancy at birth, probability of survival to age seventy-five, and
potential years of life lost) for all causes of death and selected causes, and
they examined time trends in both rates and rate ratios between the top and
bottom income quintiles. The time period starts after the introduction of
publicly financed hospital and physician services in Canada (Medicare) but
before policymakers were exposed to ideas about nonmedical determinants of
health, and it ends during a period of public-sector retrenchment.

The mortality pattern over time is largely consistent with what we expect would
have happened if ideas about nonmedical determinants of health had been acted
upon in Canada. Between 1971 and 1996 mortality rates dropped for all income
quintiles, for both sexes and for most causes of death (with the exception of
breast cancer for women, prostate cancer for men, and injuries from motor
vehicle accidents for both sexes). Mortality differentials between rich and poor
generally persisted over time, although the differentials tended to become
smaller, particularly for women. Declines in mortality rates and inequality
across income groups continued steadily over the period for some causes of
death, occurred very early in the study period for others, and accelerated in
recent years for still others.

But the story is not entirely one of progress. Mortality rates were higher and
inequality greater over time for a few causes of death: lung cancer for women
and mental disorders, AIDS, and ill-defined conditions for both sexes. These
changes for the worse tended to occur over the past five to ten years, a time
when health care spending in Canada was falling, the
prevalence of low income was rising, and unemployment rates were dropping. These
factors can hardly account for  changes in lung cancer rates for women, however,
a finding for which one must look back several decades for an explanation.
Sorting out the causes of changes over time is a complex undertaking and one for
which comparing similar data across multiple countries would only be a partial
solution.

Influence On Nonhealth Policy Making

Now, let us step back from examining changes in health to examine the influence
of ideas about nonmedical determinants of health on policy making outside the
health sector, a link in the causal chain that would have to be established to
confirm that any changes in Canadians' health are attributable to (and not just
consistent with) action on the basis of these ideas. The power of these ideas
comes from their potential to influence the development of nonhealth public
policies by adding health to the list of possible outcomes when policy
alternatives are being evaluated.9 The test of the influence of these ideas on
policy making would therefore be to identify whether these ideas have played a
role at the margin, for example, by informing a decision in the finance sector
that otherwise would have been informed by an assessment of economic and
political considerations.

The only Canadian case-study research on this issue has found that ideas about
the nonmedical determinants of health have not had a powerful influence on
either agenda setting or policy development outside the health sector." of the
five cases of policy change that were cited as having been informed by these
ideas, only two cases involved policy changes outside health, and only one of
these policy changes was influenced by these ideas. But this case-the decision
to pool the human services budgets in Prince Edward Island, justified in part on
the grounds that such a change would facilitate cross-sectoral resource
allocations in fine with the determinants of health-turned out to constitute a
symbolic use of these ideas." Ideas about nonmedical determinants of health did
not influence the policy decision per se but instead were invoked to build
support for a decision made for other reasons.

Many Canadian policymakers and researchers would argue that the National
Children's Agenda was informed by ideas
about nonmedical determinants of health, particularly research on early
childhood development.2 This policy development process constitutes an excellent
example of how policymakers and researchers can work closely together to
understand a problem's causes and the options for addressing it. No one has yet
examined the case in a systematic way, however, to ascertain whether research on
early childhood development was influential in the policy-making process and, if
so, through what mechanisms and why." At best we can say that the actions were
consistent with these ideas and offer hypotheses about why the ideas may have
been influential.4

Presence Of Ideas In PolicyMaking Environments

We now examine whether and how ideas about nonmedical determinants of health
have permeated the highest levels of Canadian government. After all, perfectly
good ideas can be trumped by many other considerations. For example, who wins,
who loses, and by how much, as well as assessments of the administrative
capacities of government departments, can influence the identification and
selection of policy alternatives. Surveys of legislators' and policy advisers'
awareness of and attitudes toward these ideas provide a reasonable test of
whether the ideas are reaching people who can act on them and what they think
about them, with the caveat that their responses may be affected by a social
desirability bias (that is, they may tell us what they think we want to hear).

In a recent survey of Canadian federal and provincial policy advisers in
finance, labor, social services, and health, our
preliminary findings suggest that advisers in labor and social services saw
health as a relevant outcome for their sector." More than 80 percent of them
felt that they should consider health determinants in all government
initiatives, and (quite surprisingly given that one would expect economic and
social outcomes to trump health outcomes in those two sectors) more than 60
percent of them disagreed with the statement that there are more important
considerations in policy making than a policy's impact on health. More than half
of these advisers were not willing to see their responsibilities for improving
health relegated to health departments.

Policy advisers in finance were another matter. Ideas about nonmedical
determinants have largely failed to permeate this sector of Canadian government,
where much of their potential lies. These departments establish tax-and-transfer
policies, influence the broader macroeconomic environment that in turn
influences labor markets, and hold the financial reins on other departments'
spending. The finance advisers who completed surveys (and are therefore more
likely to be interested in these ideas) were much less convinced than their
counterparts in other sectors were that health determinants should be considered
in all major government initiatives. Also, they were much less aware of research
on the impact of specific nonmedical determinants on health, and much less
likely to support investments in mo;e- reso:rch or in: Policy action.16

Clarity Of The Messages For Policymakers

Now we step back one last time to examine the final link in the causal chain
that would have to be established to confirm
that any changes in the health of Canadians are attributable to action on the
basis of ideas about nonmedical determinants of health. Here we come to the
clarity of the messages for policymakers and the social groups that seek to
influence them. Past research on the role of ideas in policy making has tended
to focus on bodies of research knowledge from which clear messages have been
derived for policymakers, such as Keynesian ideas about recessions and the
benefits of countercyclical demand management." The test of the clarity of
messages about nonmedical determinants of health therefore would be whether
researchers could specify the underlying causes of problems and the policy
options to address them.

Unfortunately, the messages coming out of research on nonmedical determinants of
health are anything but clear. For one thing, these ideas have evolved over
time. For example, "health promotion-the dominant version in the 1980s-tends to
advocate a more bottom-up approach to change, emphasizing public policy as only
one of five strategies to improving health. "Population health"-the dominant
version in the 1990s-tends to adopt a more top-down approach to change, often
emphasizing public policy at the expense of other options. This observation
casts doubt on our ability to assess the impact of ideas about nonmedical
determinants of health on the health of Canadians over the past twenty-five
years. No single body of ideas existed over that time period in Canada.

Moreover, while we have a great deal of evidence about the associations between
poverty and health, unemployment and health, and social isolation and health, we
have few evaluations of how specific policies that address these three
conditions affect health and the other economic and social outcomes that matter
to us. Without this information, we cannot inform trade-offs outside the health
sector. In our survey of Canadian federal and provincial policy advisers,
respondents from all sectors agreed on one area-what is now needed to act on
these ideas: More than 80 percent indicated that they need more information
about effective interventions." To put this another way, they need information
about the health consequences of the policy alternatives that their departments
face. This observation also casts doubt on our ability to assess the impact of
ideas about nonmedical determinants of health. No body of actionable messages
exists.

Weak Links In The Causal Chain

It is extremely unlikely that changes in the level and distribution of health in
Canada during 1971-1996 can be attributable to action on the basis of ideas
about nonmedical determinants of health. The links in a causal chain that
connects these ideas to the health of Canadians are far too weak. We have no
evidence that these ideas have influenced policy making outside the health
sector. We do have evidence that these ideas are absent from or not well
regarded in finance departments, an important policy-making environment outside
the health sector. And we have lacked a stable body of actionable messages for
policymakers and the groups in society that seek to influence them.

If nonmedical factors such as income, employment, and social support can explain
a substantial part of the variation in the level and distribution of health in
populations, then research should begin to focus on the health consequences of
tax-and-transfer policies, labor market policies, and social service policies.
If it did, perhaps the findings from such research could help to inform
trade-offs between policy alternatives. A minimum wage policy may be chosen over
a tax-and-- transfer policy, for example, if both policies are expected to
achieve desired economic objectives but the former leads to better health
outcomes. in the absence of such research, these ideas have remained
marginalized in Canada, neither relevant to the health sector where a few of
their policy implications lie nor sufficiently well developed to apply beyond
the health sector where most of their implications do lie.

However, a reorientation in the trajectory of Canadian research programs on the
nonmedical determinants of health may not happen spontaneously, at least if the
past twentyfive years is any indication. Canadian health policymakers will need
to either "fish or cut bait" on these ideas. Fishing will involve selectively
funding policy-relevant research that can inform trade-offs in nonhealth
sectors, for example, by paying for health outcomes to be added to the list of
outcomes being considered when policy evaluations are being conducted in
nonhealth sectors. Cutting bait will. involve getting back to their core area of
responsibility: the health care system. Researchers also face a choice: Either
they can rethink the questions they are asking and seek to produce and transfer
ideas relevant to the margins of nonhealth sectors, or they can continue to see
their ideas marginalized.

[Footnote]
1. M. Lalonde, A New Perspective on the Health of Canadians (Ottawa: Minister of
Supply and Services Canada, 1974). I use the term ideas, as opposed to research,
to reflect that policy advisers are more likely to encounter research in the
form of ideas than in the form of specific studies. See, for example, C.H.
Weiss, "Policy Research: Data, Ideas, or Arguments?' in Social Sciences and
Modern States: National Experiences and Theoretical Crossroads, ed. P. Wagner
et al. (Cambridge: Cambridge University Press, 1991), 307-332.
2. I focus here on the social determinants of health rather than the physical
determinants or public health prevention efforts. See M. Benzeval and K. judge,
"Income and Health: The Time Dimension," Social Science and Medicine 52 (2001):
1371-1390;J.N. Lavis et al., "Work-Related Population Health Indicators,"
Canadian Journal of Public Health 92, no. 1 (2001): 72-78; and L.F. Berkman,
"The Role of Social Relations in Health Promotion," Psychosomatic
Medicine 57 (1995): 245-254. A  helpful next step would be to study the factors
that may explain patterns of action and inaction in addressing the social
determinants of health, the physical
determinants of health, and more traditional causes of morbidity and mortality
such as motor vehicle accidents and smoking.
3. R.G. Evans, M.L. Barer, and T.R. Marmor, Why Are Some People Healthy and
Others Not? The Determinants of the Health of Populations (New York: Aldine de
Gruyter, 1994).
4. Lalonde, A New Perspective on the Health of Canadians; World Health
Organization, Ottawa Charter for Health Promotion (Geneva: WHO, 1986); J. Epp,
Achieving Health for All: A Framework for Health Promotion (Ottawa: Minister of
Supply and Services Canada, 1986); Federal, Provincial, and Territorial Advisory
Committee on Population Health, Strategies for Population Health: Investing in
the Health of Canadians (Ottawa: Minister of Supply and Services
Canada, 1994); and R. Gerstein et al., Nurturing Health: A Framework on the
Determinants of Health (Toronto: Premier's Council on Health Strategy, 1991).

5. D. Black et al., Inequalities in Health: The Black Report and the Health
Divide (New York: Penguin Books, 1990); and N.E. Adler et al., "Socioeconomic
Status and Health: The Challenge of the Gradient," American Psychologist 49, no.
1 (1994): 15-24. Language matters, however, and some of the changes in the
language invoked in Canada have masked substantive differences in intent and
meaning. See A. Robertson, "Shifting Discourses on Health in Canada: From Health
Promotion to Population Health," Health Promotion International 13, no. 2
(1998):155-166.
6. Federal, Provincial, and Territorial Advisory Committee on Population Health,
Toward a Healthy Future: Second Report on the Health of Canadians (Ottawa:
Minister of Public Works and Government Services Canada, 1999).
7. J.N. Lavis and TJ. Sullivan. "Governing Health," in Market Limits in Health
Reform: Public Success, Private Failure, ed. D. Drache and TJ. Sullivan (London:
Routledge, 1999), 312-328. While policymakers at the municipal or regional
level, businesses, and community-based organizations can affect some nonmedical
determinants of health, federal and state/provincial governments clearly have
the most potential to affect them.
8. R. Wilkins, E. Ng, and J.M. Berthelot, "Trends in Mortality by Income in
Urban Canada from 1971 to 1996" (Paper presented at the Population Association
of America Annual Meeting, Washington, D.C., 29-31 March 2001).
9. Psychologists have long recognized decisionmakers' tendency to restrict the
number of consequences considered. This failure to examine multiple consequences
simultaneously, often called nonconsequentialism, is pervasive be
cause of the cognitive difficulty required in undertaking such complex
considerations. E. Shafir and A. Tversky, "Thinking through Uncertainty:
Nonconsequential Reasoning and Choice," Cognitive Psychology 24, no. 4 (1992):
449-474.
10. J.N. Lavis, "Ideas, Policy Learning, and Policy Change: The
Determinants-of-Health Synthesis in Canada and the United Kingdom," CHEPA
Working Paper no. 98-6 (Hamilton, Ont.: McMaster University Centre for Health
Economics and Policy Analysis, 1998).
11. The symbolic use of ideas can be contrasted with both conceptual and
instrumental uses. D.C. Pelz, "Some Expanded Perspectives on Use of Social
Science in Public Policy," in Major Social Issues: A Multidisciplinary View, ed.
J.M. Yinger and S.J. Cutler (New York: Free Press, 1978), 346-357.
12. C. Hertzman, "The Case for Child Development as a Determinant of Health,-
Canadian Journal of Public Health 89, supp. 1 (1998): S14-S19; and D.P. Keating
and C. Hertzman, Developmental Health and the Wealth of Nations: Social,
Biological, and Educational Dynamics (New York: Guildford Press, 1999).
13. J.N. Lavis et al., "Examining the Role of Health Services Research in Public
Policy Making," Milbank Quarterly 80, no. 1 (2002): 125-154.
14. Some hypotheses about why research about early childhood development became
influential in the policy-making process include the strength of the evidence
base, the clarity of the messages for policymakers (some intervention studies
exist), the linkage between these messages and catchy political rhetoric ("by
investing in our kids, we're investing in our future economic growth"), and the
existence of a broadbased coalition that supported the change.
15. J.N. Lavis et al., "Do Canadian Policy Advisers Care about Health?" (Paper
presented at the International Conference on the Scientific Basis of Health
Services, Sydney, Australia, 23 September 2001).
16. Some hypotheses about why policy advisers in labor and social services are
more aware of and disposed toward ideas about nonmedical determinants of health
include the synergies between economic and health outcomes in their sectors,
their general focus on the well-being of workforces and clients as opposed to
the well-being of the economy, and that knowledge-transfer efforts have been
more effectively targeted at them.
17. P.A. Hall, ed., The Political Power of Economic Ideas: Keynesianism across
Nations (Princeton, NJ.: Princeton University Press, 1989).
18. Lav,is et al., "Do Canadian Policy Advisers Care about Health?"
[Author note]
John Lavis, a physician, holds a Canada Research Chair in Knowledge Transfer and
Uptake at McMaster University in Hamilton, Ontario. He also is a Liberty Health
Scholar at the Canadian Institute for Advanced Research in Toronto.