Table 1 Four illustrative examples of evidence-to-policy discordance in the context of health policy in the Unites States

Example 1: Adolescent pregnancy

Primary: Adolescents

Secondary: Children born to adolescents; grandparents of these children

Many systematic reviews of variable quality; United States and international studies

Wide range of interventions

Mostly inconclusive and conflicting results except for use of contraceptive methods Primary studies of behavioural interventions at high risk of bias, most emphasising self-reported behavioural and ‘knowledge’ outcomes

Review conclusions often overstate intervention impact; little evidence of any intervention efficacy in improving biomedical outcomes

High quality systematic reviews only support the promotion of contraceptive use combined with education

Insufficient evidence to conclude that any youth sexual risk reduction intervention effectively prevents pregnancy

Substantial low-quality evidence exists to suggest that currently recommended interventions have no impact on pregnancy rates

United States government agencies (CDC, OAH) provide millions of dollars in domestic funding for youth sexual risk reduction programmes that have been shown to be ineffective

Agencies claim population-level ‘impact’ in reducing teen pregnancy, which is far more likely due to secular trends

OAH classify these interventions as ‘evidence based’, but criteria for this determination are very weak and methodologically unsound

Human resource potential (losses to future workforce when adolescents become parents)

Personal development potential (education, career and other life goals may be less attainable when adolescents become parents; children born to adolescent parents may live in poverty; grandparents of children born to adolescent parents may take responsibility for raising them)

Costs (social services funding of prenatal, postnatal and child healthcare; social services funding for low-income families; government and non-profit funding of programmes that do not work)

Federal government agencies’ use of obsolete and flawed method for evidence synthesis and mischaracterisation of the term ‘evidence based’ to support interventions that have little or no impact on health

Unconscious biases induced by, for instance, widespread perception that current interventions are effective in reducing pregnancy, and unwillingness to change that perspective (status quo bias)

Inability of policy-makers at all levels to discern that funded interventions are ineffective, swayed by enthusiastic statements of programme implementers and perhaps by their own career bias

Unconscious biases, such as widespread perception that current interventions are effective in reducing pregnancy, and unwillingness to change that perspective (status quo bias)

Example 2: Breast cancer

Primary: Adult Women (aged > 40)

Secondary: Adult men

Multiple systematic reviews exist including by Cochrane Collaboration [80] and USPSTF [145]. There are at least 8 clinical trials and several observational studies of mixed quality

Mammography for breast cancer screening is considered an evidence-based strategy for women aged 40–70; however, appropriateness of its recommendation is debated

At the federal level, there are programmes providing free mammography to women without access Health insurance companies are required to cover mammography cost for women > 40 years 1–2 times a year

Recommendations failed to properly assess the balance between benefit and harms at the population level High quality systematic reviews do show small benefit due to mammography and only in older age groups (50–70) Significant harms associated with the procedure, specially overdiagnosis and overtreatment outweigh benefit

Assuming 30% risk of overdiagnosis with mammography and overtreatment, for every 2000 women participating in screening, over 10 years’ time span, 1 death will be prevented but 10 healthy women will be unnecessarily diagnosed and treated

In addition to its cost, mammography has several other side effects including anxiety, pain, risk of further tests and biopsies in false positive cases and risk associated with repeated exposure to x-ray

Role of the advocacy groups that may gain professionally by disseminating false information about the excessive benefit and negligible harms associated with the mammography

Lack of knowledge of stakeholders about evidence-based principles in order to help them fully understand nuanced issues around balancing benefits and harms of such intervention

Example 3: Childhood obesity

Primary: Children and adolescents

Secondary: General population, as this generation grows older

Many systematic reviews included studies from United States and other developed countries Systematic reviews have summarised evidence by intervention content (diet, physical activity, combination, etc.), setting (school, home, community, combination, etc.), and level (policy and environmental vs. individual)

Significant heterogeneity in study measurements, specific content of interventions, and risk of bias

Lack of data about certain populations and sex limited external validity of findings

Multiple interventions promoted as evidence based by major public health entities

Lack of a comprehensive and multifaceted national level legislation to address the magnitude of the public health problem

Health and economic consequence of childhood obesity is overwhelming and continues to rise

Childhood obesity is a multifaceted phenomenon, caused by inter-linked cultural, economic, and health and general literacy barriers; however, most interventions promoted as evidence-based tend to apply a bio-medical model and single approach model with possible short-term effect under ideal circumstances with limited applicability outside of tested settings

Lobby of food industry since structural level interventions would require major regulatory and restrictive changes of their procedures and practices

Complexity of subject matter prohibits non-scientific stakeholders to be able to fully understand what works

Example 4: HIV epidemic among people who inject drugs

Primary: PWID

Secondary: sexual partners of HIV-positive individuals and their children

Current: Numerous high quality systematic reviews established the intervention effectiveness in the United States and international settings

A systematic review of economic evaluations also supports that SSPs are cost-effective

Historical: As early as 1993, several reports and reviews, conducted or commissioned by public health authorities, including the CDC and National Academy of Science, concluded that SSP is safe and reduces HIV transmission [146]

Early programmes evaluated in the United States were in Tacoma, WA and Portland, OR [147]

Provision of free sterile needles and other drug injection paraphernalia in various forms such as SSPs

This intervention is often combined with other preventive measures such as condoms, peer education and HIV testing and counselling

The evidence-based recommendation has not been fully adopted by the United States federal government as a policy

Local and state health departments as well as private donors have funded > 220 programmes in more than 37 states

To respond to re-emergence of HIV and hepatitis epidemics among PWID, the Consolidated Appropriations Act of 2015 includes language providing states and local communities, under limited circumstances, the opportunity to use federal funds to support certain components of SSPs, but not to pay for needles and syringes

No comprehensive assessment at the federal level Modelling studies conducted in different cities have shown that thousands of infected cases could have been averted if SSPs were implemented

The argument that the provision of free needles may increase drug use and injection

In the absence of RCTs, evidence from other studies was dismissed, and lack of knowledge of the principle of evidence-based medicine may play a role in how some policy-makers understood the evidence

HIV epidemic among people who inject drugs coincided with ‘war on drugs’ policy in response to the increased violence associated with crack cocaine use in the 1980s; 47 states classify syringes as drug paraphernalia, making them illegal to buy or own without a prescription

Legislators from states with more conservative constituents do not want to be seen as ‘soft’ on drug enforcement or flip-flopping on a matter of moral importance