Submited on: 26 Feb 2013 06:58:57 PM GMT
Published on: 28 Feb 2013 07:05:24 AM GMT
Methods Section Needs Work
Posted by Ms. Kailyn Conner on 02 Nov 2017 01:37:38 AM GMT Reviewed by Interested Peers

  • What are the main claims of the paper and how important are they?

    The main claims of the paper are that maternal and child health outcomes during adolescent pregnancies are impacted by partner support. The importance of these claims is not mentioned within the article, but from a public health perspective, this presents an interesting claim to women's health, since, as the paper notes, the care provided to the mother determines both the maternal and child health outcomes associated with the pregnancy.

  • Are these claims novel? If not, please specify papers that weaken the claims to the originality of this one.

    No, these claims are not novel. A 2014 study by Shah et. al. published is very similar in nature to this one, but utilizes an American population rather than a Canadian one.


    Shah MK, Gee RE, Theall KP. Partner support and impact on birth outcomes among teen pregnancies in the United States. J Pediatr Adolesc Gynecol. 2014 Feb; 27(1):14-19. doi: 10.1016/j.jpag.2013.08.002.

  • Are the claims properly placed in the context of the previous literature?

    Yes, in the context of the previous literature, both cited within the paper as well as the paper referenced above, the claims seem to adhere to the generally accepted viewpoint that partner support is associated with better maternal and child outcomes.

    Another supporting paper:

    Stapleton LR, Schetter CD, Westling E, Rini C, Glynn LM, et. al. Perceived partner support predicts lower maternal and infant distress. J Fam Psychol. 2012 Jun; 26(3): 453-463. doi: 10.1037/a0028332.


    It is contrary to one study among African-American populations.

    Straughen JK, Caldwell CH, Young AA Jr., Misra DP. Partner support in a cohort of African American families and its influence on pregnancy outcomes and prenatal health behaviors. BMC Pregnancy Childbirth. 2013 Oct 17; 13: 187. doi: 10.1186/1471-2393-13-187. 

  • Do the results support the claims? If not, what other evidence is required?

    No, the results do not currently support the claims, because the exact outcomes, how those outcomes are defined, and the exposure definition is not defined within the context of the paper. This paper requires a more detailed methods section before the results can be verified to support the claims made thereafter. Furthermore, the proper statistics need to be reported in the tables in the supplement.

  • If a protocol is provided, for example for a randomized controlled trial, are there any important deviations from it? If so, have the authors explained adequately why the deviations occurred?

    The researchers did not explicitly state the protocol for this particular study. Given that the authors stated the existence of an earlier study utilizing this methodology, I assumed this was a cohort study in which the population was chosen based on an exposure and followed to an outcome. The methodology in this paper needs bolstering; it is too scarce to provide any details as to what the researchers were analyzing and what variables and the measurement thereof were available for analysis.

  • Is the methodology valid? Does the paper offer enough details of its methodology that its experiments or its analyses could be reproduced?

    The authors do not explicitly define their outcomes or their statistical methods in the methods section. You learn the statistical tools as you parse through the results section, which is problematic for the reader. Furthermore, the methods section fails to explicitly state which statistical software was used to analyze the results and the supplemental tables frequently report the wrong test statistic for interpretation. Logistic regression analyses should always report odds ratios and confidence intervals; simply listing the associated variables without providing context does not aid the reader.


    A blunter statement of the population is also needed. Who are you considering an adolescent? What was your beginning population? What was your sample population?


    This paper could not easily be reproduced, though it does provide an in-depth analysis of possible confounders in analyzing associations between partner support and maternal health outcomes.

  • Would any other experiments or additional information improve the paper? How much better would the paper be if this extra work was done, and how difficult would such work be to do, or to provide?

    The methods section needs bolstering, and some of the information provided in the results and discussion needs to be moved to the methods. The authors also need to check their statistics and report the more meaningful results from those regression analyses. Much of the analysis of risk factors could also be minimized with the utilization of a DAG.


    I also would reconsider beginning a paper on women's health with a statement comparing the beliefs of psychologists and veterinarians.


    With this extra work, this could be a much stronger paper than it presently is, though many of the claims in the conclusion should also be reevaluated, as they make large jumps.

  • Is this paper outstanding in its discipline? (For example, would you like to see this work presented in a seminar at your hospital or university? Do you feel these results need to be incorporated in your next general lecture on the subject?) If yes, what makes it outstanding? If not, why not?

    No, I would not consider this paper outstanding in its discipline. The claims are not novel and reinforce generally held beliefs, and a similar paper in another population has already been released. The difficulty in determining the methods also presents a challenge toward calling this paper outstanding.

  • Other Comments:

    Thank you for tackling such an important issue!


    Mostly the methods section here needs work. What is your exact exposure? How was that measured? If it was a survey question, what question was asked? What were the possible responses? What was the reference group? What was your outcomes? How were they evaluated? You performed a logistic regression - I am assuming the outcomes were dichotomous? How were they measured? What defined an outcome occurring? All these questions are vital to methods sections, regardless of the study design.


    While I applaud the careful assessment of associated risk factors, a DAG would drastically minimize the work needed to cover everything here. Perhaps that is something that could be looked into?


    Furthermore, proper statistics being reported are vital to the utility of the paper you write. Logistic regression should always report odds ratios and confidence intervals; these are usable and vital pieces of information, as it defines the exact increase in odds of an event occuring given an outcome. You said you also performed a Cox regression - how was that evaluated? Where is your Kaplan-Meier table? Where are those hazard ratios? Those results were not provided to the reader.


    I also strongly suggest rethinking the use of animal behavior comparisons in a paper on women's health.

  • Competing interests:
  • Invited by the author to review this article? :
  • Have you previously published on this or a similar topic?:
  • References:


  • Experience and credentials in the specific area of science:

    Master of Public Health, concentrating in Epidemiology

  • How to cite:  Conner K .Methods Section Needs Work[Review of the article 'How Partner Support of an Adolescent Affects Her Pregnancy Outcome ' by Sheils C].WebmedCentral 2017;8(11):WMCRW003383
1 2 3 4 5 6 7 8 9
Report abuse