Maternal distress and childhood asthma

Maternal distress and childhood asthma

by | Sujit Rathod -
Number of replies: 9

From The New York Times

1. What study design is this?

2. How was the exposure defined? Comment the limitations of this measurement.

3. Calculate and interpret this: 362/4231

4.  How was the outcome defined?

"After controlling for age, smoking during pregnancy, body mass index, a history of asthma and other factors.."

5. Why did the researchers control for these? Would you have controlled for different variables?

"...they found that maternal depression and anxiety during pregnancy was significantly associated with both diagnoses of asthma and poorer lung function in their children. There was no association between childhood asthma and parents’ psychological distress in the years after pregnancy, and no association with paternal psychological stress at any time."

6. Why should this kind of language make an epidemiologist cringe?

7. Can you design an RCT to test this hypothesis?

In reply to | Sujit Rathod

Re: Maternal distress and childhood asthma

by | FATHIMA MINISHA -
Hello... I found it :-)

1) This was a prospective cohort study... They first documented the exposure- which is psychological stress during pregnancy and a few months after. They then followed the kids at 10 yrs of age to look for any history or current evidence of asthma. So going from exposure to outcome, and in real time. So its a prospective cohort study.

2) The exposure was defined as psychological stress based on a questionnaire completed in second trimester of pregnancy, and 2, 6 months after birth and after 3 years. They used a validated questionnaire, with definitions for what constitutes psychological stress. The main limitation is that its subjective and self reported- hence can result in a lot of variation in the reporting of symptoms.

3) 362/4231- this is number of mothers with psychological stress/ total number of mothers in the sample
The result will be - 8.6%
Interpretation- 8.6% of the mothers questioned showed evidence of psychological stress.
Its written 4231 children who were assessed at 10 years... however they do state in the study that they excluded twins- therefore it implies these were 4231 mothers.

4) The outcome was defined as parents reporting asthma in the child, along with objective outcome s in the form of lung function tests

5) The authors controlled for these factors because all of those factors can be confounders. On reading the paper, we can know that they have accounted for many more factors like breastfeeding, day care attendance etc. The child's home environment is another variable I would be interested in.

6) Dont know actually. Maybe because the language is complicated and not easy to read- can be confusing even. I had to read it atleast 3 times to understand it completely.

7) Here the hypothesis is that psychological stress that occur during pregnancy is associated with development of asthma in the kids. It would be unethical to have an RCT with this exposure- I mean we cannot expose pregnant women to undue psychological stress just to see if their kids develop asthma... That was be completely unethical... So I think this hypothesis can be tested only with an observational study.

Fathima
In reply to | FATHIMA MINISHA

Re: Maternal distress and childhood asthma

by | Alice Kociejowski -

Hi Fatima,

You've inspired me to have a go at this. I was very excited to see this forum because I am totally guilty of writing in the same way as the newspaper journalist. I hope by  learning 'how not to do it' I will improve my basic epi skills!

My attempt below:

1. What study design is this? Observational prospective cohort study

2. How was the exposure defined? Comment the limitations of this measurement.

Exposure – psychological distress of either biological parent during pregnancy, and at 3 years, and psychological distress of the mother at 2 and 6 months post partum.

Parents self reported, which is subject to bias. Also, it is not clear if both parents (and child, post partum) were in the same household during the pregnancy and follow up, which might influence the 'association' between asthma and psychological stress.

3. Calculate and interpret this: 362/4231 = 0.086 or 8.6%. 362 mothers of 4231 children born (8.6% of children's mothers surveyed)  reported psychological distress during pregnancy in at least one of multiple surveys conducted.

4.  How was the outcome defined? The outcome was ‘ever having been diagnosed with asthma’ according to the parents, plus the outcome of an FEV1 lung function test.

"After controlling for age, smoking during pregnancy, body mass index, a history of asthma and other factors.."

5. Why did the researchers control for these? Would you have controlled for different variables?

Family history – asthma, eczema, hayfever

Place of residence – air pollution, mould, pets

Household smoking (after pregnancy)

"...they found that maternal depression and anxiety during pregnancy was significantly associated with both diagnoses of asthma and poorer lung function in their children. There was no association between childhood asthma and parents’ psychological distress in the years after pregnancy, and no association with paternal psychological stress at any time."

6. Why should this kind of language make an epidemiologist cringe?

They have not explained their research and findings in terms of  Chance, Bias and Confounding. Therefore it is unclear if the findings are ‘significant’ in any way and they cannot draw any conclusions as to causality from their findings.

7. Can you design an RCT to test this hypothesis?

It would be ethically and morally challenging to make a pregnant woman  (or her partner) distressed and equally difficult to prevent a pregnant woman from getting distressed in order to understand causality! Though I would be a willing candidate for the latter cohort...

I also wonder about the use of the FEV1 test to 'check' an asthma diagnosis - asthma is usually defined as 'reversible airways disease' and I think would require administration of a drug, salbutamol, to check for reversibility. There are lots of other clinical considerations as well around relying on parents to confirm/ deny an asthma diagnosis.






In reply to | FATHIMA MINISHA

Re: Maternal distress and childhood asthma

by | Sujit Rathod -
Fathima - glad you found your way over. :)
In reply to | Sujit Rathod

Re: Maternal distress and childhood asthma

by Benguimbis | BENEDICT PIERRET TAA NGUIMBIS ESSEME -

I read only the press release and here are my takes for 1-2-3-7

1-2-7  I think they intended to do a Cohort study since the outcome was measured years later. If I wanted to do a hypothetical RCT we would have randomized depression and anxiety status. This brings to the identification of Immortal time bias in the study. Instead of comparing incident status, they have compared the prevalent status. Moreover, the exposure is time-varying and the exposure status of an individual could change during the study form exposed to unexposed leading to the misalignment of person-time from exposed to unexposed status. The exposure is also susceptible to measurement bias since the exposure is self-reported and subjective, misclassification of the exposure is likely. On the other hand, the exposure is susceptible to selection bias (immortal time bias), a better definition would have been to compare a positive diagnosis of depression and anxiety within the first term of the pregnancy, and to exclude all the women with such all women with a history of depression or anxiety prior to the trial to only keep incident exposed. We cannot design a RCT for this exposure as it is neither manipulable nor ethical but designing a target trial can solve a lot of problems. 

3. From the answer above 362/4231 is the prevalence of the exposure in the whole study 

In reply to | Sujit Rathod

Re: Maternal distress and childhood asthma

by | Olivia El Jassar -
1. prospective cohort study

2. exposure = 'psychological stress' during 2nd trimester, 2 and 6mo after giving birth, and 3 years later
Psychological stress was assessed via a questionnaire - highly subjective, not standardised, may over/underestimate their stress

3. 0.0855 = 8.6% prevalence of psychological distress

4. outcome = a diagnosis of asthma amongst children by 10 years old (researchers also assessed poorer lung function in general via FEV)

5. important confounding factors i.e. all are major risk factors for children getting asthma; by controlling for them, you aim to eliminate their influence, allowing you to focus in on the impact 'psychological stress' alone on asthma outcome

6. no measures given? Ie. no measure of outcome effect, exposure effect, p-values etc??

7. Tricky to randomise groups of psychologically stressed mothers into 2 groups and probably major ethical issues (children getting asthma unnecessarily when this could have potentially been prevented by treating their mother's stress.)
In theory you could get a random sample of 'psychologically stressed' women, randomise them to a treatment (anxiety/depression medication) and placebo group. Assess along the way the impact of treatment. Evaluate child lung function and asthma status after X years.
In reply to | Sujit Rathod

Re: Maternal distress and childhood asthma

by | Yasmin Sonbol -
1. What study design is this?
This is a cohort study.

2. How was the exposure defined? Comment the limitations of this measurement.
Exposure defined based on a hypothesis that pregnant parents with psychological distress might increase risk for asthmatic children, and data was generated through a questionnaire on psychological stress in the second trimester of pregnancy, and again three years later. Then again two and six months after giving birth.

3. Calculate and interpret this: 362/4231
Of the total number of parents (4231 including fathers and mothers), there is 8.56% prevalence of mothers with psychological distress during pregnancy

4. How was the outcome defined?
When the children were 10 years old, parents reported whether their child had ever been diagnosed with asthma.

"After controlling for age, smoking during pregnancy, body mass index, a history of asthma and other factors.."
5. Why did the researchers control for these? Would you have controlled for different variables?
Yes, I might have also included familial medical background in regards to asthma, and any other factors that could cause asthma like the presence of pets at home, for instance.

"...they found that maternal depression and anxiety during pregnancy was significantly associated with both diagnoses of asthma and poorer lung function in their children. There was no association between childhood asthma and parents’ psychological distress in the years after pregnancy, and no association with paternal psychological stress at any time."
6. Why should this kind of language make an epidemiologist cringe?
Maybe because it denies the association while it was previously stated that they developed asthma as an effect from psychological distress, so it might be a bit confusing as which is true.

7. Can you design an RCT to test this hypothesis?
No, because it seems that the exposure is during the uterus as the study mentioned and the outcome appears later in the lifetime of the child (after 10 years as per the study)
So taking a random sample at a random time will not be the best case scenario to help in generalizing the results.
In reply to | Sujit Rathod

Re: Maternal distress and childhood asthma

by | JUDITH MARGARET BURCHARDT -
1. What study design is this?
Observational cohort study

2. How was the exposure defined? Comment the limitations of this measurement.
Self assessed mood using a questionnaire
This is not an objective measurement and we do not know if it reproducible

3. Calculate and interpret this: 362/4231
8.6% point prevalence of self assessed maternal ‘clinically significant psychological distress’

4. How was the outcome defined?
In two ways - by parental history and by Fev.

"After controlling for age, smoking during pregnancy, body mass index, a history of asthma and other factors.."

5. Why did the researchers control for these? Would you have controlled for different variables?

These are potential confounders. I would have controlled for family history of atopy and if possible exercise in the children. It is possible that children with depressed mothers have less exercise and so have poorer Fev

"...they found that maternal depression and anxiety during pregnancy was significantly associated with both diagnoses of asthma and poorer lung function in their children. There was no association between childhood asthma and parents’ psychological distress in the years after pregnancy, and no association with paternal psychological stress at any time."

6. Why should this kind of language make an epidemiologist cringe?

Lack of definition? Need to specify what the confidence intervals are to say statistically significant and also the degree of association which is defined as not significant.

7. Can you design an RCT to test this hypothesis?

No, not possible to randomise people and then make some distressed as discussed by others.
In reply to | JUDITH MARGARET BURCHARDT

Re: Maternal distress and childhood asthma

by | OLGA VIACHESLAVOVNA KOZHAEVA -
Hi all

some thoughts where different or additional to the above.

2. How was the exposure defined? Comment the limitations of this measurement.

From article: “Well-validated questionnaires - psychological stress in the second trimester of pregnancy, and again three years later. The mothers also completed questionnaires at two and six months after giving birth.”

- I am not sure if self-report per se is an issue as this seems to be the only way to assess mental well-being? so long as the questionnaire is well validated - as stated in the article, I suppose that it could generate valid answers. do you have other thoughts?

- The fact that exposure was only measured in a specific trimester is a limitation – there is potential for measurement error if this extrapolated to the entire pregnancy period

- Unclear how exactly depression and anxiety were defined – was this based on a score? It seems vague since the authors state that psychological distress was measured – what level of psychosocial distress indicated depression and anxiety? Was the main outcome depression or anxiety, or psychosocial distress? all of these are used interchangeably in the article and leads to lack of clarity as  to what exactly the exposure is.

4. How was the outcome defined?

Parent-reported ever asthma diagnosis in the first 10 years of the child’s life supplemented by results of FEV test.
- Unclear from article whether outcome was defined as present if both measures indicated asthma or only one of these. In addition, FEV cut-off point was not specified. so at least based on the article, there is lack of clarity on outcome definition, as well.

5. "After controlling for age, smoking during pregnancy, body mass index, a history of asthma and other factors.."
Why did the researchers control for these? Would you have controlled for different variables?

These factors were probably found to be associated with childhood asthma in prior research and could have been confounders if also associated with depression during pregnancy and not on causal pathway. Age is a priori confounder in addition. I would also measured and controlled for over-crowding/social class as a proxy for that, as social class may be independently associated with maternal depression and asthma but is not on causal pathway.

6. "...they found that maternal depression and anxiety during pregnancy was significantly associated with both diagnoses of asthma and poorer lung function in their children. There was no association between childhood asthma and parents’ psychological distress in the years after pregnancy, and no association with paternal psychological stress at any time."
Why should this kind of language make an epidemiologist cringe?

"Significantly associated" points to the use of potentially arbitrary cut-off point on the probability scale. It would have been preferable to state ‘there was strong statistical evidence (p=x)“.. and also “no statistical evidence of association (p=x)” instead of just “no association”.
Accessibility
Reset All

Background Colour

Font Face

Font Size

1

Text Colour