Epi in the News

Dear Sujit,

Here are my answers:

1. What is the study design?
- It is a historical cohort study.    

2. What are the exposure and outcome?
- The exposure is the sugar tax, and the outcome is the incidence rates of hospital admissions for carious tooth extraction in children under 18

2b. What is the mechanism of effect?
The mechanism of effect involves reducing the development of caries through the reduction of sugar intake, which is a known risk factor for dental caries

Overall there has been a drop of 3.7 admissions per 100,000 zero to 18-year-olds since 2018.

3. Is the 3.7 an incidence or prevalence figure? Is it a difference or a ratio?
- 3.7 is a drop in incidence which is a difference (the reduction in hospital admissions)

4. Can you come up with (epidemiologic) alternative explanations for this drop?
- Another possible explanation might be the change in oral hygiene practices in children. However It's unlikely that substantial changes would have occurred in such a short time frame.

5. Do we need a trial to prove causation? What would this involve? Since we already have a very strong biological plausibility for causality, as sugar is known to be in the causal pathway of caries (as well as in other diseases), a RCT is probably not needed here. Moreover this would pose some ethical issues as it would lack equipoise.

Thanks for the interesting article. Here are some of my replies::

1. What is the study design?
- Historical cohort study. Records were examine at a point in time and exposures/outcomes were elicited for the time period.

2. What are the exposure and outcome?
Exposure: tax on sugar in soft drinks, and its impact on children aged 18 and below
Outcome: number of Children requiring tooth extraction under general anaesthetic in hospital

2b. What is the mechanism of effect?
The concept is that sugar sweetened drinks are a large source of sugar consumption in children. By taxing the amount of sugar in the drinks, beverage makers are incentivized to reduce the amount of sugar in their drinks to keep the price constant. Or the price is transferred to the customer, and customers are less willing to pay for drinks with a lot of sugar. High sugar consumption is associated with increased rates of tooth decay in children, so reducing sugar consumption is likely to reduce rates of tooth decay, of which extractions under GA in hospital is a proxy for.

Overall there has been a drop of 3.7 admissions per 100,000 zero to 18-year-olds since 2018.
3. Is the 3.7 an incidence or prevalence figure? Is it a difference or a ratio?
Without reading the original paper and understanding the health system better, it's hard to answer. If the hospital records are reliable for recording all cases of tooth extractions under GA in children, then it is a prevalence figure. But if it does not include a segment of the population (e.g. only accounts for public hospital admissions, not private), then it is technically a sample of the population and an incidence figure.
This is a difference figure.
They also reported this difference is a 12% drop, that puts the change in context.

4. Can you come up with (epidemiologic) alternative explanations for this drop?
- other improvements in dental hygiene. E.g. changes in floridation of the water, more regular teeth brushing and flossing due to dental hygiene education in schools, improved access to preventive dental care in the community, etc.
- Shifting of extractions away from GA in hospital to outpatient setting under LA. This is big. I'd guess that most extractions in children under 18 are done in the outpatient setting. We could be missing an overall increase in teeth extractions. Parents may be opting away from GA extractions for reasons such as the increased risks (especially to adolescents)
- If the information does not include private hospitals, then the shift could be towards GA extractions done in private instead of public hospitals. Dental care in UK is privatized I believe?

5. Do we need a trial to prove causation? What would this involve? What is the study design?
Causation that a sugar tax will decrease the number of extractions done under GA? No. Because I think that question is not what we're interested in answering. 

The outcome we're most interested in is probably sugar consumption. As Marco mentioned, we already have good evidence that reducing sugar intake will probably reduce dental caries.

 
What we probably want is an evaluation study on the effectiveness of implementing the sugar tax: Is the sugar tax a good intervention for reducing harmful outcomes of excessive sugar consumption in children? I feel the answer to this question is probably yes based on Coherence, but that alone may not be enough to "prove" causality. In reality, we will have to rely on imperfect outcomes to evaluate the effectiveness of the policy. I think being slightly uncertain about the causality is fine. 

In a perfect world, we could do a cluster-based randomised control trial. With comparable geographic areas implementing the sugar tax vs without the tax. With a reliable method of measuring the incidence of dental caries in each cluster (perhaps through the collection of dental records), we could come up with a comparison of incidence of dental caries in areas with vs without a sugar tax before and after intervention. This would hopefully fulfil the temporality, consistency, and experimental aspects of the Bradford Hill Criteria. How feasible it is to conduct such a study is another matter. 

Good afternoon Sujit, I hope you had a good weekend.
Let´s have a look to your questions:

1.What is the study design? It is an observational descriptive population based cross sectional study.

2. What are the exposure and outcome? the exposure is the sugar, and the outcome is teeth fall

2b. What is the mechanism of effect?

Overall there has been a drop of 3.7 admissions per 100,000 zero to 18-year-olds since 2018.
I think they are talking about the incide risk

3. Is the 3.7 an incidence or prevalence figure? Is it a difference or a ratio? I think we are talking about an incidence risk, and it is a difference as they say there has been a drop of 3.7 pero 10.000 since 2018, and not per person/year

4. Can you come up with (epidemiologic) alternative explanations for this drop? better care on teeth hygine could explain the situation, there is also a lobby against sugar even before the law started

5. Do we need a trial to prove causation? What would this involve? Well in this case basic research could safe children from a trial, as you can test exposure to sugar but If I did a trial:
- Controls: children with a low sugar diet
-Intervention group: children with no fat sugar diet, (only the recommended by the who)
The outcome could be tooth decay, but a better outcome to study is obesity, type two diabetes.

Hi Sujit,

The following are my answers:

1. This is a retrospective cohort study where the yearly hospital admission rates (after the sugar tax came into force in April 2018) due to carious teeth extraction under general anaesthesia among children aged 18 and under was compared with that before the soft drinks levy (SDIL) implemented.

2. The exposure was implementation of the soft drinks levy (SDIL), and the outcome was incidence rates of hospital admissions due to carious teeth extraction under general anaesthesia.

2b. The mechanism of effect is reductions in development of dental caries in children aged 18 and under due to reduced sugar intake after the implementation of the soft drinks levy, and thereby reductions in hospital admissions due to carious teeth extraction under general anaesthesia.

3. 3.7 is an incidence figure. It is a difference.

4. Alternative (epidemiologic) explanations for this drop can be: (i) increased awareness regarding harmful effects of sweetened soft drinks among general population after the implementation of the soft drinks levy, (ii) change in clinical practice (extraction being done in outpatient settings with local anaesthesia), (iii) improved oral hygiene practices, (iv) fluoridation of drinking water, (v) change in formulations of soft drinks, etc.

5. Trial is not needed to prove causation, as previous studies showed strong evidence of an association between sugar intake and development of carious teeth. Therefore, it is not needed to prove that reduced sugar intake causes reductions in hospital admissions due to carious teeth extraction under general anaesthesia.
Moreover, such a trial involves ethical issues.

Thanks!
Madhutandra