Mednotes

L5 Public Health What can we do to prevent disease_

Properties2
tagspublichealth, lecture
aliasesL5 Public Health What can we do to prevent disease_

16 min read

Overview

This lecture uses cardiovascular disease (CVD) as a worked example of the public health approach to disease prevention. It follows a research-to-action cycle from measuring the size of the problem, through identifying causes, to developing and disseminating interventions, and reviews the global and Aotearoa/NZ burden of CVD and its ethnic inequities, the evidence linking cholesterol and blood pressure to IHD mortality, relative vs absolute risk, the Framingham risk score, the current NZ CVD risk assessment tool and treatment targets, and the strengths and limitations of risk-scoring approaches.

The public health research-to-action framework

Three epidemiological questions map onto the stages of a cyclical Public Health: Research & Action process, which loops back on itself via ongoing evaluation and monitoring:

  • How common is the disease, and does it vary by person, place, or time? (descriptive epidemiology) corresponds to Defining and Measuring the Problem.
  • Is the disease more, less, or about as common among people “exposed” to a factor versus “not exposed”? (analytical epidemiology) corresponds to Describing Causes & Consequences.
  • Does an intervention reduce disease occurrence or improve survival? (analytical epidemiology) corresponds to Developing and Evaluating Interventions.
  • The cycle continues to Disseminating Effective Policy and Practice, then feeds into ongoing evaluation and monitoring, which loops back to the start.

Terminology

  • Acute Coronary Syndrome (ACS): applied to patients in whom there is suspicion or confirmation of acute myocardial ischaemia or infarction. Three traditional types: non-ST-elevation MI (NSTEMI), ST-elevation MI (STEMI), and unstable angina.
  • Related terms used for the same disease group: Ischaemic Heart Disease, Coronary Artery Disease.
  • Stroke is divided into Ischaemic and Haemorrhagic types.

Global burden of CVD

  • Number of CVD deaths rose steadily from about 6 million (1990) to about 9 to 10 million (2019) in both females and males, tracking closely with males slightly higher.
  • China has by far the highest and fastest-rising number of CVD deaths of any country (about 2.5 million in 2010 rising to about 4.5 million in 2019), followed by India; the Russian Federation, USA and Indonesia are substantially lower.
  • Proportion of CVD deaths by cause (2019): ischaemic heart disease 49.2%, ischaemic stroke 17.7%, intracerebral haemorrhage 15.5%, hypertensive heart disease 6.2%, subarachnoid haemorrhage 2%, cardiomyopathy/myocarditis 1.8%, atrial fibrillation/flutter 1.7%, rheumatic heart disease 1.6%, other cardiovascular/circulatory diseases 1.5%, non-rheumatic valvular heart disease 0.9%, aortic aneurysm 0.9%, peripheral artery disease 0.4%, endocarditis 0.4%.
  • Age-standardised CVD DALY rates vary by region: Central Asia, Eastern Europe and Oceania have among the highest rates (roughly 8,000 to 14,000 per 100,000), with males generally higher than females in most regions; High-income Asia Pacific and High-income North America have the lowest rates (roughly 1,500 to 2,500 per 100,000).

CVD burden in Aotearoa/NZ and ethnic inequities

  • ACS is a leading cause of morbidity and mortality in NZ.
  • Leading causes of death in 2018 (all population): cancer (30.3% of all deaths), ischaemic heart disease (14.9%), cerebrovascular disease (7.4%).
  • For Māori, the leading causes of death in 2016 were cancer, ischaemic heart disease and chronic lower respiratory disease.
  • In NZ deaths data (2021), IHD and stroke occupy the largest share of any cause category, ahead of lung cancer, colorectal cancer, lower respiratory infections and other cancers.
  • In the equivalent NZ DALYs (2021) breakdown, IHD and stroke are still large but relatively smaller than in the deaths data, because mental health and musculoskeletal conditions (back pain, other musculoskeletal disorders, osteoarthritis, depression, anxiety, headaches) contribute disproportionately more to disability burden than to deaths.
  • IHD mortality (age-standardised, per 100,000) rose from about 100 in 1948 to a peak of about 250 around 1968 to 1975, then declined steadily to below 50 by 2021. Cerebrovascular disease mortality followed a similar but lower pattern, peaking at about 120 around 1970 to 1975 and declining to below 25 by 2021.
  • 2021 IHD deaths and age-standardised rates per 100,000 by ethnicity (all sexes): Asian 186 (24.5), European/Other 3664 (38.9), Māori 528 (72.5), Pacific 216 (70.1); overall 4594 deaths (rate 42.9). Māori and Pacific mortality rates are roughly double those of European/Other and Asian groups, in both sexes.
  • Over 2006/07 to 2014/15 (ages 35 to 84), age-standardised IHD hospitalisation and death rates declined over time for all ethnic groups in both women and men, but Māori consistently had the highest rates throughout, followed by Pacific, then Indian, with European and Other Asian the lowest; men’s rates were substantially higher than women’s overall.

Causes and consequences: identifying risk factors

  • Epidemiological studies lead to identification of risk factors; named examples: Framingham Heart Study, Seven Countries Study, Nurses Health Study.
  • Risk factors are classified as modifiable or unmodifiable.
  • Trials are then used to test interventions targeting these risk factors.

.

Cholesterol and blood pressure as risk factors for IHD mortality

  • A meta-analysis of individual data from 61 prospective studies (55,000 vascular deaths, Prospective Studies Collaboration) found that the total/HDL cholesterol ratio was the strongest predictor of IHD mortality, more than twice as informative as total cholesterol alone.
  • IHD mortality hazard ratio rises steeply and roughly log-linearly as usual total cholesterol increases, within every 10-year age band from 40 to 89; for a given cholesterol level, the relative mortality risk increase is much greater at younger ages than at older ages.
  • The same pattern holds for blood pressure: IHD mortality rises steeply and roughly log-linearly with increasing usual systolic and diastolic blood pressure within each age band, and for a given blood pressure the hazard is higher in younger age groups.
  • A prospective meta-analysis of 90,056 participants in 14 randomised statin trials (Cholesterol Treatment Trialists’ Collaboration) found the proportional benefit of LDL-lowering on major coronary events per 1.0 mmol/L LDL reduction was consistent across baseline prognostic subgroups (previous disease, age, sex, treated hypertension, diabetes history, diastolic BP, total/LDL/HDL cholesterol, triglycerides), with little heterogeneity by presence or level of other risk factors. Overall rate ratio 0.79 (95% CI 0.76 to 0.80): 7.4% events with treatment vs 9.8% with control.

Relative risk and absolute risk

  • Relative risk is the ratio of the probability of an event (e.g. developing a disease) occurring in an exposed group (e.g. smokers) to the probability of it occurring in a comparison, non-exposed group (e.g. non-smokers).
  • Absolute risk is an individual’s risk of developing a disease over a specific time period, expressed as a probability; e.g. a 1 in 10 risk of developing a disease over the next 5 years can also be expressed as a 10% risk.

Framingham risk score

  • Converts an individual’s risk factor levels into a points score, which is then used to estimate their absolute 10-year CHD risk (%).
  • Points are assigned by age band, and for total cholesterol, HDL cholesterol, systolic blood pressure, diabetes status, and smoking status (point allocations differ between men and women for some factors, e.g. HDL cholesterol and diabetes).
  • Higher point totals correspond to markedly higher absolute 10-year risk, e.g. more than 14 points corresponds to over 53% (total CHD) or over 45% (hard CHD) 10-year risk.

NZ cardiovascular disease risk assessment

  • Up to 2018, NZ guidelines used Framingham Study data and six risk factors (blood pressure, cholesterol, age, diabetes status, smoking status, sex) plotted on colour-coded charts (separate for men and women, by age band, diabetes status and smoking status) mapping systolic BP against total cholesterol:HDL ratio into 5-year CVD risk categories: low (<2.5%), mild (2.5 to 10%), moderate (10 to 20%), high (20 to 30%), very high (>30%).
  • The current guideline (Cardiovascular Disease Risk Assessment and Management for Primary Care, Ministry of Health, updated February 2018) estimates 5-year CVD risk, based on NZ cohort data from the PREDICT study (400,728 patients in the NZ GP database, aged 45 to 74 years for men and 55 to 74 years for women, from 30/40 years for Māori, Pacific and South Asian people). Recommendations are evidence-based and consistent with other NZ and international guidelines.
  • Recommended ages to start risk assessment (Table 4): people without known risk factors, men from 45 and women from 55; Māori, Pacific or South Asian people, men from 30 and women from 40; people with other known cardiovascular risk factors or high diabetes risk, men from 35 and women from 45 (risk factors include family history of diabetes, family history of early heart attack/stroke death or hospitalisation before age 50, familial hypercholesterolaemia, smoking, gestational diabetes, HbA1c 41 to 49 mmol/mol, BMI ≥30 or truncal obesity, eGFR <60 but >45 mL/min/1.73m², atrial fibrillation); people with diabetes (type 1 or 2), from time of diagnosis for both sexes; people with severe mental illness, from age 25 for both sexes. Established CVD is automatically classed as high risk.
  • Variables used in the current NZ risk assessment tool: age, gender, ethnicity, NZ Index of Deprivation, family history of premature CVD or type 2 diabetes, past history of CVD or familial hypercholesterolaemia, history of diabetes (including its duration and renal function), smoking status, HbA1c, blood pressure, non-fasting lipids, eGFR (renal function), BMI, and medications.

Treatment decisions and targets

  • Recommended interventions by risk level (Table 5) all include lifestyle advice (diet, weight management, physical activity, smoking cessation) at every level, with drug therapy escalating by risk: established CVD, strong evidence supports pharmacotherapy for modifiable risk factors plus antiplatelet therapy for secondary prevention, reviewed annually; >15% CVD risk, strong evidence supports statins and BP lowering, reviewed annually; 5 to 15% CVD risk, discuss the magnitude of benefit of statins/BP lowering (higher risk means more likely to benefit), repeat risk assessment at 5 years (5 to 9% risk) or 2 years (10 to 14% risk); <5% CVD risk, medication has limited benefit, repeat assessment at 10 years (<3% risk) or 5 years (3 to 5% risk). People with diabetes are reviewed annually regardless of risk level.
  • Lipid target: drug treatment if total cholesterol/HDL ratio ≥8 regardless of CVD risk; if CVD risk ≥15%, target LDL below 1.8 mmol/L.
  • Blood pressure target: drug treatment if persistent office BP ≥160 mmHg systolic and/or ≥100 mmHg diastolic regardless of CVD risk; target below 130/80 mmHg office BP (caution in older people).
  • Diabetes target: HbA1c 50 to 55 mmol/mol for younger and fitter patients, or 55 to 64 mmol/mol for older, co-morbid and frail patients.
  • Repeat CVD risk assessment intervals: <3% risk, repeat at 10 years; 3 to 9% risk, repeat at 5 years; 10 to 14% risk, repeat at 2 years; ≥15% risk, repeat at 1 year as part of annual management review.
  • Risk-assessment algorithms should be applied with additional clinical judgement, rather than relied on alone, in: people over 74 years, those with serious mental illness, morbidly obese patients, those with chronic kidney disease, people taking medicine for HIV, and those with autoimmune and/or systemic inflammatory disorders.

Strengths and limitations of risk scoring

  • WHO (2007) argues the strength of risk scoring is that it gives a rational way of targeting intervention and making best use of resources; alternative approaches focused on single risk factors label a very large, mostly incorrect, segment of the population as high risk. Risk scoring shifts focus from managing individual risk factors to reducing an individual’s overall risk, matching intervention intensity to total risk.
  • Limitations (Jackson, Kerr & Wells, 2013): about 15 to 20% of patients who develop CVD have none of the traditional risk factors and would be labelled low risk by current prediction algorithms; atherosclerosis is increasingly recognised as beginning in childhood and associated with the same traditional risk factors, so intervening only in adulthood may be too late; increasing life expectancy makes 5 or 10-year risk estimated at age 50 less clinically relevant, given actual CVD risk over a further 40 to 50 years of remaining life; many algorithms are derived from late 20th century cohort studies, when CVD rates were much higher than now.

Self-test

  1. Describe the three-question framework (descriptive and analytical epidemiology) used to define the problem, describe causes, and evaluate interventions, and state which stage of the public health research-to-action cycle each question corresponds to.
  2. Define acute coronary syndrome and name its three traditional types.
  3. What did the treemap comparison of NZ deaths versus NZ DALYs (2021) show about the relative contribution of IHD/stroke compared with mental health and musculoskeletal conditions, and why does that difference arise?
  4. Describe the overall trend in NZ ischaemic heart disease and cerebrovascular disease mortality rates from 1948 to 2021.
  5. Distinguish the pattern of 2021 IHD mortality rates between Māori/Pacific and European/Other/Asian ethnic groups.
  6. List the three named epidemiological studies credited with identifying cardiovascular risk factors, and distinguish modifiable from unmodifiable risk factors as a classification.
  7. According to the Prospective Studies Collaboration meta-analysis, which single lipid measure was the strongest predictor of IHD mortality, and how does the relative mortality risk from a given cholesterol or blood pressure level differ between younger and older age bands?
  8. What did the Cholesterol Treatment Trialists’ Collaboration meta-analysis of 90,056 participants find about how the proportional benefit of LDL-lowering on coronary events varied across baseline risk-factor subgroups?
  9. Distinguish relative risk from absolute risk, using the smoking example and the “1 in 10” example given.
  10. Describe how the Framingham risk score converts an individual’s risk factor levels into an estimate of absolute 10-year CHD risk.
  11. List the six risk factors used in the pre-2018 NZ Framingham-based CVD risk charts.
  12. Describe the population the current NZ PREDICT-based CVD risk assessment tool is derived from, and explain why the recommended starting age for risk assessment differs for Māori, Pacific and South Asian people compared with people without known risk factors.
  13. List the variables used in the current NZ CVD risk assessment tool.
  14. Describe how recommended lifestyle advice, drug therapy and follow-up interval change across the four CVD risk categories (established CVD, >15%, 5-15%, <5%).
  15. State the lipid, blood pressure and HbA1c treatment targets given for cardiovascular risk management.
  16. Which patient groups should have CVD risk-assessment algorithms applied with additional clinical judgement rather than relied on alone?
  17. According to Jackson, Kerr and Wells (2013), what are the four main limitations of current cardiovascular risk-scoring algorithms?
  18. Explain, using the WHO (2007) quotation, why a risk-scoring approach is considered preferable to an approach focused on managing single risk factors individually.

Answers

Graph View

Backlinks