Letters

Depression and unintended pregnancy in young women

BMJ 2002; 324 doi: https://doi.org/10.1136/bmj.324.7345.1097 (Published 04 May 2002) Cite this as: BMJ 2002;324:1097

Paper raises at least three concerns

  1. Robert S Kahn, assistant professor (robert.kahn@chmcc.org)
  1. Division of General Pediatrics, Children's Hospital Medical Center, Cincinnati, OH 45229, USA
  2. University of South Dakota School of Medicine, Sioux Falls, SD 57105, USA
  3. Ipas, Chapel Hill, NC 27516, USA
  4. 11542 Clear Creek Drive, Pensecola, FL 32514-9703, USA
  5. Elliot Institute, PO Box 7348, Springfield, IL 62792-7348, USA

    EDITOR—Reardon and Cougle's paper raises at least three concerns.1

    Firstly, their analyses do not address the stated hypothesis. No results indicate whether “prior psychological state is equally predictive of subsequent depression among women … regardless of whether they abort or carry to term.” Nevertheless, their unstated hypothesis, focused on abortion and depressive symptoms, may be the more central question.

    Secondly, the final sample of women is surprisingly small. Only 421 of the initial 4463 women reported a first abortion or first unintended delivery between 1980 and 1992. Is it possible that the question in 1992 asking pregnancy intention actually referred to a much narrower time frame (that is, a delivery between the biannual surveys)? Little information is given about the abortion question; it is possible that the index unintended pregnancy defined in 1992 resulted in neither the first abortion nor the first delivery.

    Thirdly, the discussion omits mention of possible residual confounding. The national longitudinal survey of youth (NLSY) uses a four item abbreviated version (NLSY Cronbach α 0.35) of Rotter's original 60 item locus of control scale, which itself is probably an inadequate proxy for prior psychological state. Furthermore, a one year measure of income may be only a modest proxy for a person's lifetime socioeconomic position. 2 3 The robustness of the authors' findings could be examined with other available measures. For example, the 1980 Rosenberg self …

    View Full Text

    Sign in

    Log in through your institution

    Subscribe