Skip to main content

Table 3 Estimated impact of reduction strategies on TB risk factor prevalence

From: Social and behavioral risk reduction strategies for tuberculosis prevention in Canadian Inuit communities: a cost-effectiveness analysis

Background change in prevalence of risk factor without reduction strategya Estimate of effect of reduction strategy Reduction strategy’s annual reach Change in prevalence of risk factor with reduction strategyb
 Decrease in number of smokers by 0.013% per year A systematic review indicates that persons receiving pharmacotherapy and counselling in communities receiving mass media interventions were twice as likely to quit commercial tobacco smoking than those who received usual care; Relative Risk = 2.36 (95% CI 1.01–5.50) [55]. A SimSmoke simulation notes that 25% price increase leads to 7% reduction in smoking prevalence within three years and increases over time to 14% [51, 56] The entire smoking population is exposed to taxation plus the availability of pharmacotherapy and counselling (but only a fraction make quit attempts; see Table 4). Mass media campaigns, however, reach both smokers and non-smokers. All components of this combined strategy occur within the first year of the simulation, and do not repeat after that. Their effect, however, lasts longer [23]. Decreases the smoking population by 5.54% per year for the first 3 years, then by 1.01% per year for the next 7 years. Returns to background decrease of 0.013% per year for the remaining 10 years.
Increase in number of heavy drinkers by 0.07% per year A cohort study in California exploring the effectiveness of a holistic cultural treatment program suggests that inpatient participants experience a 28.3% absolute reduction in use of alcohol and drugs in the 30 days following their treatment session (31.3% at baseline vs. 3.0% at follow-up) [54]. The mobile treatment center serves 8 individuals per session × 4 sessions × 3 regions = 96 individuals. In the model, this is scaled up by a factor of 5 (480 individuals ≈ 7% of the heavy drinking population). The 20 sessions in each region occur within the first year of the model, and do not repeat after that. As such, their effect also occurs within the first year of the simulation. Decreases the number of heavy drinkers by 1.91% in the first year. Returns to background increase of 0.07% per year for the remaining 19 years.
Decrease in number of persons living with food insecurity by 0.36% per year A case-control study conducted among 533 households in 14 communities in Northern Manitoba noted that the presence of a country food program increases the likelihood of being food secure; Odds Ratio = 20.64 (95% CI 2.42–176.08) [53]. Because of the uncertainty around the point estimate and its magnitude, we have used the lower bound of the 95% confidence interval as a conservative measure of impact, and investigated its effect further in sensitivity analysis. All four initiatives in this strategy have been scaled up to reach the entire food-insecure population. The Canada Prenatal Nutrition Program, however, reaches only food-insecure pregnant women (≈2.3% of the food insecure population). Decrease in the number of food insecure persons by 0.87% each year for all 20 years.
Increase in persons living in overcrowded housing by 2.27–5.76% per yearc The cohort study that our overcrowding reduction strategy is based on reports a 46% absolute reduction in overcrowding among those who were accommodated by the new housing units (65.5% at baseline vs. 19.5% at follow-up) [50]. 426 housing units built × 3.3 individuals on average in each unit at follow-up = 1406 individuals (≈ 6% of the overcrowded population). The housing units are built in the first year of the model, but their associated maintenance recurs annually. The corresponding effect on overcrowding prevalence occurs within the first year of the simulation. Decrease in persons living in overcrowded housing by 0.49% for the first year, reffecting this effect superimposed on background increase in overcrowding. Returns to background increase by 2.27–5.76% each year for the next 9 years, to reach 100% by year 10. No change in prevalence for the remaining 10 yearsd.
  1. aThese changes were informed by trends in TB risk factor prevalence over time in Nunavut
  2. bCalculations illusrating how the estimates of effect translate into changes in prevalence are provided in Additional file 1 - Supplemental Table 4
  3. cOvercrowding prevalence increases parabolically, not linearly, so the rate at which it increases does not remain constant (but remains between 2.27 and 5.76%)
  4. dOvercrowding prevalence reaches 100% during the 10th cycle of the simulation, because of background population growth exceeding housing supply, so there is no additional increase after that