3) The author's response by Morten Frisch

312

survey-based-cross-sectional study in Denmark. Int J Epidemiol 2011;40:1367–81. Barros AJ, Hirakata VN. Alternatives for logistic regression in cross-sectional studies: an empirical comparison of models that directly estimate the prevalence ratio. BMC Med Res Methodol 2003;3:21. Savu A, Liu Q, Yasui Y. Estimation of relative risk and prevalence ratio. Stat Med 2010;29:2269–81. Morris BJ. Why circumcision is a biomedical imperative for the 21st century. BioEssays 2007;29:1147–58. Krieger JN, Mehta SD, Bailey RC et al. Adult male circumcision: effects on sexual function and sexual satisfaction in Kisumu, Kenya. J Sex Med 2008;5: 2610–22. Kigozi G, Watya S, Polis CB et al. The effect of male circumcision on sexual satisfaction and function, results from a randomized trial of male circumcision for human immunodeficiency virus prevention, Rakai, Uganda. BJU Int 2008;101:65–70. Waskett JH, Morris BJ. Fine-touch pressure thresholds in the adult penis. (Critique of Sorrells ML et al. BJU Int 2007;99:864–69.) BJU Int 2007;99:1551–52. Mao LM, Templeton DJ, Crawford J et al. Does circumcision make a difference to the sexual experience of gay men? Findings from the Health in Men (HIM) Cohort. J Sex Med 2008;5:2557–61. Yang L, Ruan LM, Yan ZJ et al. [Sexual function and mental state in patients with redundant prepuce or phimosis]. (Article in Chinese.) Zhonghua Nan Ke Xue 2010; 16:1095–97.

Stenram A, Malmfors G, Okmian L. Circumcision for phimosis: a follow-up study. Scand J Urol Nephrol 1986;20: 89–92. ´mezCorte´s-Gonza´lez JR, Arratia-Maqueo JA, Go Guerra LS. Does circumcision have an effect on female’s perception of sexual satisfaction? [in Spanish]. Rev Invest Clin 2008;60:227–30. Kigozi G, Lukabwe I, Kagaayi J et al. Sexual satisfaction of women partners of circumcised men in a randomized trial of male circumcision in Rakai, Uganda. BJU Int 2009;104:1698–701. Morris BJ, Gray RH, Castellsague X et al. The strong protection afforded by circumcision against cancer of the penis. (Invited Review). Adv Urol 2011; Article ID 812368:1–21. Tsen HF, Morgenstern H, Mack T et al. Risk factors for penile cancer: results of a population-based case-control study in Los Angeles County (United States). Cancer Causes Control 2001;12:267–77. Backes DM, Bleeker MC, Meijer CJ et al. Male circumcision is associated with a lower prevalence of human papillomavirus-associated penile lesions among Kenyan men. Int J Cancer 2011; [Epub 26 May 2011]. Weiss HA, Hankins CA, Dickson K. Male circumcision and risk of HIV infection in women: a systematic review and meta-analysis. Lancet Infect Dis 2009;9:669–77. doi:10.1093/ije/dyr180 Advance Access publication 28 November 2011 ß The Author 2011; all rights reserved.

Author’s Response to: Does sexual function survey in Denmark offer any support for male circumcision having an adverse effect? From MORTEN FRISCH Department of Epidemiology Research, Statens Serum Institut, DK-2300 Copenhagen S, Denmark. E-mail: mfr@ssi.dk

Novel findings in our population-based survey, which had participation rates of 48% in men and 54% (not 40%, as wrongly mentioned by Morris et al.) in women, suggest, but by no means prove, the existence of non-trivial associations of male circumcision with frequent orgasm difficulties in men and with a range of frequent sexual difficulties in women, including orgasm difficulties, dyspareunia and a sense of incomplete sexual needs fulfilment. Morris et al. should not be blamed for feeling unconvinced by our findings. However, as these critics repeatedly refer to Morris’ pro-circumcision manifesto1 as their source of knowledge, their objectivity must be questioned. Morris et al. express concern over possible overfitting in our logistic regression models because we included a number of potentially confounding variables that differed between circumcised and uncircumcised men and between women with circumcised and uncircumcised spouses. However, as seen in Tables 3–6

of our paper, models with adjustment only for age provided odds ratios (ORs) similar to those obtained in the fully adjusted model, suggesting that this is mostly a theoretical concern. Next, Morris et al. suggest that we should have corrected for multiple testing even though such statistical manoeuvres are, at best, unnecessary and, at worst, deleterious to sound statistical inference in most epidemiological studies.2 Morris et al. also claim that prevalence ratios would have been more appropriate measures of association than ORs. However, despite Morris et al.’s firm statement to the contrary, there is nothing inherently inappropriate about using ORs in cross-sectional studies, even in situations with common outcomes. In such situations, however, ORs should not be misinterpreted as prevalence ratios. We would have been wrong to claim that our OR of 3.26 implied that frequent sexual difficulties were 3.26 times more common in women with

Downloaded from http://ije.oxfordjournals.org/ at Statens Serum Institute on March 19, 2012

INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

LETTERS TO THE EDITOR

to declare their conflicts of interest. Even in situations that are out of context, Morris promotes himself as a neutral ‘authority on the extensive medical benefits of this simple surgical procedure’,8 whereas at the same time he argues that neonatal male circumcision ‘should be made compulsory’ and that ‘any parents not wanting their child circumcised really need good talking to’.9 In contrast, we conducted our survey without setting up any a priori hypotheses, because the limited and inconclusive literature on possible sexual consequences of circumcision would permit almost any imaginable a priori hypothesis. We had no intent to prove an already known ‘truth’ or disprove its contradiction. It is ironic that Morris et al. question the credibility of our findings, postulating that I have an ‘active involvement in opposition to male circumcision’. I have never expressed any objection, ethical, medical or other, against male circumcision as such. Unlike Morris, who believes that ‘circumcision is a biomedical imperative for the 21st century’,1 I could not care less whether fully informed, healthy adults choose to get circumcised or not. Likewise, when foreskin pathology is present (which does not include the physiological tightness of the foreskin experienced transiently by most boys), and the problem cannot be treated conservatively, preputioplasty or partial circumcision may be a relevant solution, even in minors and others who are unable to consent to the operation. However, because ethical discussions about ritual circumcision are sometimes distorted by strong personal views, I openly declared that I have participated in national debates over ethical issues surrounding male and female circumcision. Like in critical letters to the editor following other recent studies that failed to support their agenda,10–12 Morris et al. air a series of harsh criticisms against our study. As seen, however, the points raised are not well founded. It seems that the main purpose, as with prior letters, is to be able in future writings to refer to our study as an ‘outlier study’ or one that has been ‘debunked’, ‘rejected by credible researchers’ or ‘shown wrong in subsequent proper statistical analysis’. This in spite of the fact that our study was carried out using conventional epidemiological and statistical methods, underwent peer-review and was published in an international top-ranking epidemiology journal. I would like to thank the IJE editors for withstanding the pressure from one particularly discourteous and bullying reviewer who went to extremes to prevent our study from being published. After the paper’s online publication, I have received emails from colleagues around the world who felt our contribution was useful and potentially important. One colleague informed me that the angry reviewer was the first author of the above letter to the editor. In an email, Morris had called people on his mailing list to arms against our study, openly admitting that he was the reviewer and that he had tried to get the paper

Downloaded from http://ije.oxfordjournals.org/ at Statens Serum Institute on March 19, 2012

circumcised spouses than in women with uncircumcised spouses. Nowhere in our paper did we interpret ORs in such a flawed manner. In accordance with the cited reference3 we simply noted that frequent sexual difficulties were more common in women with circumcised spouses and that the associated fully adjusted OR was 3.26. Next, Morris et al. argue that our finding of considerably higher rates of frequent orgasm difficulties in (partially) circumcised than uncircumcised Danish men (11 vs 4%, OR ¼ 3.26) may not apply in countries where circumcision means complete amputation of the foreskin. This may well be the case. If partial amputation of the foreskin truly entails frequent orgasm difficulties in a noticeable proportion of men (as experienced by 11% of circumcised men in our study), comparable proportions may well be larger and associated ORs even higher in countries where circumcised men experience greater tissue loss due to more extensive circumcision procedures. Obviously, more data are needed from rigorous studies using carefully constructed questionnaires. The questionnaires used to assess potential sexual problems in the two cited randomized controlled trials in Kenya and Uganda were not presented in detail in the original publications.4,5 Rather than blindly accepting such findings as any more trustworthy than other findings in the literature, it should be recalled that a strong study design, such as a randomized controlled trial, does not offset the need for high-quality questionnaires. Having obtained the questionnaires from the authors (RH Gray and RC Bailey, personal communication), I am not surprised that these studies provided little evidence of a link between circumcision and various sexual difficulties.4,5 Several questions were too vague to capture possible differences between circumcised and not-yet circumcised participants (e.g. lack of a clear distinction between intercourse and masturbation-related sexual problems and no distinction between premature ejaculation and trouble or inability to reach orgasm). Thus, non-differential misclassification of sexual outcomes in these African trials probably favoured the null hypothesis of no difference, whether an association was truly present or not. Morris et al. should be commended for their creative attempt to dismiss the higher prevalence of frequent dyspareunia in women with circumcised (12%) than uncircumcised (4%) spouses (ORs between 4.17 and 9.00). They suggest that Danish women with circumcised spouses may be so psychologically troubled by the shape of their spouse’s penis that it might result in painful intercourse. A more plausible explanation would be that reduced penile sensitivity may raise the need among some circumcised men for more vigorous and, to some women, painful stimulation during intercourse in their pursuit of orgasm. Two of the authors, Morris and Waskett, both internationally recognized circumcision activists,6,7 forget

313

314

INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

References 1

Morris BJ. Why circumcision is a biomedical imperative for the 21st century. Bioessays 2007;29:1147–58. Perneger TV. What’s wrong with Bonferroni adjustments. Br Med J 1998;316:1236–38. Barros AJ, Hirakata VN. Alternatives for logistic regression in cross-sectional studies: an empirical comparison

of models that directly estimate the prevalence ratio. BMC Med Res Methodol 2003;3:21. Krieger JN, Mehta SD, Bailey RC et al. Adult male circumcision: effects on sexual function and sexual satisfaction in Kisumu, Kenya. J Sex Med 2008;5:2610–22. Kigozi G, Watya S, Polis CB et al. The effect of male circumcision on sexual satisfaction and function, results from a randomized trial of male circumcision for human immunodeficiency virus prevention, Rakai, Uganda. BJU Int 2008;101:65–70. Circleaks.http://circleaks.org/index.php?title¼Brian_Morr is (8 August 2011, date last accessed). Circleaks.http://circleaks.org/index.php?title¼Jake_H._ Waskett (8 August 2011, date last accessed). Morris BJ. Renin, genes, and beyond: 40 years of molecular discoveries in the hypertension field. Hypertension 2011;57:538–48. YouTube.http://www.youtube.com/v/7yDvL4hNny4 (8 August 2011, date last accessed). Morris BJ, Wodak A. Circumcision survey misleading. Aust N Z J Public Health 2010;34:636–37. Waskett JH, Morris BJ, Weiss HA. Errors in meta-analysis by Van Howe. Int J STD AIDS 2009;20:216–18. Waskett JH, Morris BJ. Fine-touch pressure thresholds in the adult penis. BJU Int 2007;99:1551–52. doi:10.1093/ije/dyr181 Advance Access publication 28 November 2011 ß The Author 2011; all rights reserved.

Humans and models: converging ‘truths’ From TIM A BRUCKNER1* and CLAIRE MARGERISON-ZILKO2 1

Program in Public Health & Department of Planning, Policy, and Design, University of California, Irvine, CA, USA and Population Research Center & Center for Social Work Research, University of Texas, Austin, TX, USA

*Corresponding author. Program in Public Health & Department of Planning, Policy, and Design, University of California, 202 Social Ecology I, Irvine, CA, 92697, USA. E-mail: tim.bruckner@uci.edu

We thank Jay Kaufman for his thoughtful commentary1 regarding our recent manuscript in which we reported a positive relation between acute income gains and accidental deaths among Cherokee Indians in rural North Carolina.2 Although we agree with many of Kaufman’s points, we would like to respond to a key question that holds relevance to most analyses using time series data: how should epidemiologists approximate the counterfactual value of a population exposed at a specific point in time? One approach to deriving counterfactual values in time involves using a model-based framework. In our analysis of the Cherokee response to acute and large cash disbursements from a local Casino, we employed a Poisson regression with a conventional log-linear functional form.3 To control for confounding by temporal patterns in accidental deaths (e.g.

seasonality), we included indicator variables for calendar months and years. Identification of an effect of the Casino payments on accidental deaths, therefore, relies on a systematic deviation—in the 20 exposed months of the Casino disbursements—above expected values derived from the specific underlying (multiplicative) functional form of accidental deaths. The analyst, of course, could impose different modelbased assumptions regarding temporality of accidental deaths (e.g. additivity), yet Kaufman notes the problem of insufficient statistical power to test such assumptions. In other words, the analyst has limited ability to detect what we call ‘wrong model bias’ in the functional form of a time series. An alternative approximation of the counterfactual value, which Kaufman suggests, involves using a comparison series of accidental deaths among

Downloaded from http://ije.oxfordjournals.org/ at Statens Serum Institute on March 19, 2012

rejected. To inspire his followers, Morris had attached his two exceedingly long and aggressive reviews of our paper (12 858 words and 5291 words, respectively), calling for critical letters in abundance to the IJE editors. Breaking unwritten confidentiality and courtesy rules of the peer-review process, Morris distributed his slandering criticism of our study to people working for the same cause. Rather than resorting to such selective distribution among friends, Morris should make both reviews freely available on the internet by posting them in their entirety on his pro circumcision homepage (www.circinfo.net). Alternatively, interested readers should feel free to request them from me at the e-mail address above. Despite poorly founded criticisms and attempts at obstruction our findings suggest that male circumcision may be associated with hitherto unappreciated negative sexual consequences in a non-trivial proportion of men and women. Further carefully conducted studies are needed.