Recent eLetters

Displaying 11-20 letters out of 460 published

  1. Re: Improving the governance of patient safety in emergency care: a systematic review of interventions

    We welcome this systematic review on interventions to improve the governance of patient safety in emergency care settings. This work highlights the value of information contained in adverse events derived from medical record review and incident reporting systems. The authors identified four qualities shared by effective incident reporting systems: (i) staff education on the importance and learning purpose of reporting, (ii) multiple and constantly available reporting options, (iii) a short reporting form to minimise burden, and (iv) structural feedback by presenting descriptive statistics, findings of incident root-cause analyses and improvement actions (p. 10, [1]).

    Our experience of implementing an Emergency Medicine-specific incident reporting system for use across Australasian hospitals supports these findings. The Emergency Medicine Events Register has supported site champions at participating hospitals to educate fellow staff about incident reporting and provides an on-line (, streamlined data collection form tailored to emergency medicine, and automated feedback including descriptive summaries of incidents [2]. Reporting an incident takes just under 6 minutes. Other measures were taken to support reporting, such as the ability to claim continuing professional development points and statutory protection of the data from discovery (qualified privilege) [2].

    The Emergency Medicine Events Register is a voluntary, anonymous system that is managed by an independent third party (the Australian Patient Safety Foundation). The decision for anonymity and independent management was made to enhance system uptake by doctors. Lack of medical engagement in incident reporting is often ascribed to fear and mistrust of incident reporting systems and their misuse by hospital administrators and regulators [3] and a perception that error is inevitable and should be managed 'in-house' or by self-regulation [4]. Hesselink et al's 2016 review also supported our model, with the proviso that implementation of a non-anonymous peer review process had improved reporting rates in one study [5]. Hesselink et al (2016) concluded that "anonymity of reporting and management of incident reports by an independent third party may not be necessary if an incident reporting and peer review process is perceived to be safe" (p. 10).

    We agree with this proposition but believe that, mostly, doctors' perceptions are that incident reporting is not yet considered safe, and therefore that anonymity is an essential element to an incident reporting system. Anonymity imposes limitations on the type of feedback that can be provided - it may only be generic, such as case studies [6]. However, anonymous systems can support hospital-based incident reporting systems that have a strong focus on accountability (as opposed to system learning) and, in general, poor uptake by doctors [7,8].

    Meaningful engagement of doctors has been identified as one of three keys to unlocking the full potential of incident reporting systems [9]. Ownership of patient safety concerns by specialties, such as Emergency Medicine, is a necessary step to developing sustainable system-wide improvements. Specialty led systems that engage trainees and fellows in data collection and analysis, are integrated with professional organisations and their related activities (such as training and education) and are supported by third party patient safety organisations, can promote national and international patient safety agendas [10].

    However, the barriers to doctors reporting incidents are significant. Emergency medicine-specific barriers to incident reporting include: difficulties in training all staff, the need for one-on-one support for site champions, perceived duplication with hospital-based systems and a view that reporting was only a consultant-level activity [2]. Such barriers inform the continued development of the system. One such recent development in the Emergency Medicine Events Register has been an expansion in scope that now allows for consumer reporting [11].


    1. Hesselink G, Berben S, Beune T, et al. Improving the governance of patient safety in emergency care: a systematic review of interventions. BMJ Open, 2016;6:e009837.

    2. Schultz TJ, Crock C, Hansen K, et al. Piloting an online incident reporting system in Australasian emergency medicine. Emerg Med Australas, 2014;26:461-7.

    3. Waring JJ. Beyond blame: cultural barriers to medical incident reporting. Soc Sci Med, 2005;60:1927-35.

    4. Rosenthal M. How doctors think about medical mishaps. In: Rosenthal M, Mulcahy L, Lloyd-Bostock S, eds. Medical Mishaps. Buckingham: : Open University Press. 1999. 141-53.

    5. Reznek MA, Barton BA. Improved incident reporting following the implementation of a standardized emergency department peer review process. Int J Qual Health Care, 2014;26:278-86.

    6. Deakin A, Schultz TJ, Hansen K, et al. Diagnostic error: Missed fractures in emergency medicine. EMA - Emerg Med Australas, 2015;27:177-8.

    7. Hobgood C, Weiner B, Tamayo-Sarver JH. Medical error identification, disclosure, and reporting: Do emergency medicine provider groups differ? Acad Emerg Med, 2006;13:443-51.

    8. Rowin EJ, Lucier D, Pauker SG, et al. Does error and adverse event reporting by physicians and nurses differ? Jt Comm J Qual Patient Saf, 2008;34:537-45.

    9. Mitchell I, Schuster A, Smith K, et al. Patient safety reporting: a qualitative study of thoughts and perceptions of experts 15 years after 'To Err is Human'. BMJ Qual Saf, 2015, published online first: doi:10.1136/bmjqs-2015-004405

    10. Jones DN, Benveniste KA, Schultz TJ, et al. Establishing national medical imaging incident reporting systems: issues and challenges. J Am Coll Radiol, 2010;7:582-92.

    11. Emergency Medicine Events Register. Consumer Report. 2016. (accessed 23rd March 2016).

    Conflict of Interest:

    None declared

    Read all letters published for this article

    Submit response
  2. Critical appraisal of the article

    I would like to submit this paper to the authors of this article for the purpose of commenting on the methods and results in the form of critical appraisal:

    This study addresses a clearly focused question in terms of population, the exposure, the outcomes, and the timeline of the study. The primary hypothesis was clearly outlined in this study as well as the secondary hypothesis. The study design chosen, prospective cohort is an appropriate design for addressing the research question as the paper was attempting to study a certain exposure (parental styles) and its effect on subsequent substance use among adolescents. The cohort group may not be representative of the population, as study subjects were not randomly selected. Also, it is worth noting that this paper does not clearly state the inclusion/exclusion criteria used in recruiting the students, which may hinder the external validity of the study.

    In addition, the outcome measurements used in this paper were validated and the components of each questionnaire were clearly defined. However, given that all the outcome measures were subjective and rely heavily on self- report, the possibility of biased responses is high especially that it involves asking children about harmful behaviors as some of them would have a tendency to answer the questions based on social desirability rather than the truth. This can tremendously affect the internal validity of the results.

    Furthermore, the authors have identified and analysed several important confounding factors such as gender, parental behaviour, behaviour in school, deviant peers, delinquency, among other factors, which is important to get a genuine picture of the real effect of exposure. However, other important confounders were overlooked such as income, level of education, and whether parents are separated, which in my opinion are very important confounding factors that affect parents' attitudes towards their child's substance use habits.

    The follow up period was adequate (32 months) to determine whether baseline exposure influenced the outcome at teenage years of the students. Also, sensitivity analysis was performed to determine whether results were different for those lost to follow up. The final results of the study concluded no significant association between parenting styles and subsequent use of substance by adolescents. The paper included adjusted and non-adjusted odds ratio with 95% confidence intervals, which is very vital for interpretation of results. Nevertheless, the results may have been affected by reporting bias, and ignored confounders as mentioned earlier.

    It is interesting that results of this study are different from similar studies with larger sample sizes (1, 2). As opposed to this study, both studies reported a significant relationship between authoritative parenting style and reduction in substance use among adolescents (1, 2). Could the difference be attributed to cultural differences in parental approval of underage drinking?

    In conclusion, this study has its shares of strengths and weaknesses. A strong point in this study is its large sample size, which comes from different schools in Sweden, which raises the possibility for generalization of results. Another strong point is that the study has a long follow up time and a high response rate from adolescents and minimal dropouts, which increases the power of the study. On the other hand, weaknesses include the heavy reliance on self-report on measuring the outcomes and missing important confounding factors.


    1. Calafat A, Garcia F, Juan M, Becona E , & Fernandez-Hermida. Which parenting style is more protective against adolescent substance use? Evidence within the European context. Drug and Alcohol Dependence, 2014;138:185-192.

    2. Shakya HB, Christakis NA, & Fowler JH. Parental Influence on Substance Use in Adolescent Social Networks. Arch Pediatr Adolesc Med Archives of Pediatrics & Adolescent Medicine, 2012;166(12):1132-1139.

    Conflict of Interest:

    None declared

    Read all letters published for this article

    Submit response
  3. APOC3 testing in CHD

    In their article, Lin and coworkers have suggested that "APOC3 S2 allele carriers might be advised to go for regular checkups to reduce the risk of adverse cardiac event". The recommendation is based on a pooled analysis of association findings addressing four APOC3 variants, SstI, T- 455C, C-482T and C1100T, in coronary heart disease (CHD). While such tests have already been commercialized for CHD, major flaws in the study by Lin et al. warrant a cautionary note. Thus odds ratios and 95% confidence intervals reported for SstI deviate from the original data [1-4], numerous studies are missing [5-8] and genetic models of risk have been muddled [9]. Despite the authors' claim of adopting a rigorous methodology that conformed to guidelines, and despite a fully transparent peer review, many more errors have gone unnoticed. These comprise the failure to eliminate overlapping samples for T-455C [10 11], treating hazard ratios as odds ratios [12] and an intriguing exclusion of large data sets from the studies referenced for C-482T [12], C1100T [12] and SstI or C3175G [13]. Taken together, pending a thorough re-analysis of APOC3 association data, testing for the above variants is best postponed in a context of CHD.


    1. AshokKumar M, Subhashini NG, SaiBabu R, Ramesh A, Cherian KM, Emmanuel C. Genetic variants on apolipoprotein gene cluster influence triglycerides with a risk of coronary artery disease among Indians. Mol Biol Rep, 2010;37:521-7.

    2. Sediri Y, Kallel A, Feki M, Mourali S, Elasmi M, Abdessalem S, Mechmeche R, Jemaa R, Kaabachi N. Association of a DNA polymorphism of the apolipoprotein AI-CIII-AIV gene cluster with myocardial infarction in a Tunisian population. Eur J Intern Med, 2011;22:407-411.

    3. Abd El-Aziz TA, Mohamed RH, Hashem RM. Association of lipoprotein lipase and apolipoprotein C-III genes polymorphism with acute myocardial infarction in diabetic patients. Mol Cell Biochem, 2011;354:141-50.

    4. Chhabra S, Narang R, Lakshmy R, Vasisht S, Agarwal DP, Srivastava LM, Manchanda SC, Das N. Apolipoprotein C3 SstI polymorphism in the risk assessment of CAD. Mol Cell Biochem, 2004;259:59-66.

    5. Baroni MG, Berni A, Romeo S, Arca M, Tesorio T, Sorropago G, Di Mario U, Galton DJ. Genetic study of common variants at the Apo E, Apo AI, Apo CIII, Apo B, lipoprotein lipase (LPL) and hepatic lipase (LIPC) genes and coronary artery disease (CAD): variation in LIPC gene associates with clinical outcomes in patients with established CAD. BMC Med Genet, 2003;4:8.

    6. Masana L, Febrer G, Cavanna J, Baroni MG, Marz W, Hoffmann MM, Shine B, Galton DJ. Common genetic variants that relate to disorders of lipid transport in Spanish subjects with premature coronary artery disease. Clin Sci, 2001;100:183-90.

    7. Ding Y, Zhu MA, Wang ZX, Zhu J, Feng JB, Li DS. Associations of polymorphisms in the apolipoprotein APOA1-C3-A5 gene cluster with acute coronary syndrome. J Biomed Biotechnol, 2012;2012:509420.

    8. Paulweber B, Friedl W, Krempler F, Humphries SE, Sandhofer F. Genetic variation in the apolipoprotein AI-CIII-AIV gene cluster and coronary heart disease. Atherosclerosis, 1988;73:125-33.

    9. Bhanushali AA, Das BR. Influence of genetic variants in the apolipoprotein A5 and C3 gene on lipids, lipoproteins, and its association with coronary artery disease in Indians. J Community Genet, 2010;1:139-48.

    10. Olivieri O, Stranieri C, Bassi A, Zaia B, Girelli D, Pizzolo F, Trabetti E, Cheng S, Grow MA, Pignatti PF, Corrocher R. ApoC-III gene polymorphisms and risk of coronary artery disease. J Lipid Res, 2002;43:1450-7.

    11. Martinelli N, Trabetti E, Bassi A, Girelli D, Friso S, Pizzolo F, Sandri M, Malerba G, Pignatti PF, Corrocher R, Olivieri O. The -1131 T>C and S19W APOA5 gene polymorphisms are associated with high levels of triglycerides and apolipoprotein C-III, but not with coronary artery disease: an angiographic study. Atherosclerosis, 2007;191:409-17.

    12. Wong WM, Hawe E, Li LK, Miller GJ, Nicaud V, Pennacchio LA, Humphries SE, Talmud PJ. Apolipoprotein AIV gene variant S347 is associated with increased risk of coronary heart disease and lower plasma apolipoprotein AIV levels. Circ Res, 2003;92:969-75.

    13. Tobin MD, Braund PS, Burton PR, Thompson JR, Steeds R, Channer K, Cheng S, Lindpaintner K, Samani NJ. Genotypes and haplotypes predisposing to myocardial infarction: a multilocus case-control study. Eur Heart J, 2004;25:459-67.

    Conflict of Interest:

    None declared

    Read all letters published for this article

    Submit response
  4. Incorrect trial number; correct number is NCT00368199

    Our paper has the trial number NCT0036819, and it should read NCT00368199 (See

    Conflict of Interest:

    None declared

    Read all letters published for this article

    Submit response
  5. No evidence for kanuka honey being an effective treatment of rosacea

    Braithwaite et al. [1] claim that kanuka honey is an effective treatment of rosacea. Both their study medication as well as their study design and statistical analysis are seriously flawed and do not allow this conclusion. First, the study medication is inappropriate to show an effect of kanuka honey. The verum was not kanuka honey (as the title suggests), but a mixture of kanuka honey and 10% glycerine (Honevo). As control, an entirely different preparation was chosen, i.e. Cetomacrogol a mitigating and protective cream containing macrogol cetostearyl ether 22, cetiol, sorbitol and the preservative ascorbic acid, used for dry and sensitive skin, eczema and itching. This induced an obvious risk of bias and lack of blinding. Moreover, if the specific effects of kanuka honey (versus another honey) were to be tested, a different honey should have been tested with the same amount of glycerine. Second, the study would have been powered for a 25% response in the control and a 50% response in the Honevo group. However, the observed effect sizes were much lower (17.4% and 34,3%, respectively). Even if subsequently p-values less than 0.05 were calculated, no conclusions can be drawn due to this a priori lack of statistical power. Thus, both the title and the conclusion of the publication are not justified by the study design and data.


    1. Braithwaite I, Hunt A, Riley J, et al. Randomised controlled trial of topical kanuka honey for the treatment of rosacea. BMJ Open 2015;5:e007651. doi:10.1136/bmjopen-2015-007651

    Conflict of Interest:

    None declared

    Read all letters published for this article

    Submit response
  6. Changing the corresponding author

    We kindly ask you to note that the corresponding author should be JY Chung. He is the third author in our manuscript.

    Conflict of Interest:

    None declared

    Read all letters published for this article

    Submit response
  7. Systems science in Public Health: building capacity, navigating language... with MeSH?

    We agree with Carey et al.(1) that systems science methods are an important addition to the methodological and conceptual toolbox used to approach complex public health problems and risk factors.(2) In order to realize the full potential of these methods, it is necessary to review and critique the state of the science regarding systems science applications in the public health research context, both calling out successes and learning from failures or shortcomings. However, for such efforts to have the greatest impact, they must be done carefully and represent the body of work studied. This is difficult when language used to describe relevant methods is inconsistent, as is true of many approaches being re-imagined or emerging, including systems science.

    A recent interpretive systematic review of systems thinking and systems science by Carey et al.(1) focused on a subset of public health literature (i.e. obesity, tobacco, alcohol, and social determinants of health). Through inductive qualitative analysis, the authors described and analyzed papers and highlighted the need to expand the capability of public health researchers to engage with the full range of systems science methodologies. We welcome the contribution of Carey et al. and believe the findings, in combination with other reviews and critiques of systems science methods in public health, will enable public health researchers to better integrate systems science methods with existing methodological tools and connect them to pressing public health challenges. Unfortunately, limitations of the study itself highlight critical gaps and capacity-building needs to advance the use of systems science in public health.

    A key limitation of Carey et al.'s review is a mismatch between the broad conceptual definition of systems thinking and science articulated and the much narrower operational definition used to conduct the review. The authors note that their methods were not positioned to conduct an exhaustive review, but claim their identified articles are likely representative of systems science publications in the field. Unfortunately, we believe the mismatch between what they describe broadly as their focus of the review and their operationalization of "systems science" provides a potentially limited representation that has important implications for continued development and integration of systems science methods in public health.

    Carey et al.'s operational definition, as indicated by the search query presented, underspecifies systems science methods. System dynamics, agent-based modeling (ABM), and network analysis are often identified as three prevalent systems science methods in public health research.(3) The search terms used to identify publications included the first of these methods, system dynamics, but not ABM or network analysis. This likely overestimates the proportion of systems science studies that apply system dynamics and underestimates the proportion of studies that use other methods. While several instances of these unnamed methods were identified, likely through the broader terms included in the search (e.g. 'systems science'), a more robust search is likely to have captured many more studies using these methods. This oversight is especially troubling given that ABM and network analysis studies qualified for inclusion in the study, but high profile (e.g. a manuscript by Christakis and Fowler assessing the social contagion of obesity through social networks (4)) and relevant studies (e.g. a manuscript by Orr using ABM to study racial disparities in obesity (5)) were not identified and included in the review.

    Complicating matters further, a more inclusive review might explicitly name a variety of other systems science methods, including soft systems methodology, critical systems heuristics, community based operations research, and discrete event simulation. (6,7,8) In our experience searching this literature, calls for understanding problems as shaped by systems often include the broader systems terms used in Carey et al.'s search, but applications of specific and relevant systems science methods are less likely to do so as their focus is more on describing the application and implications of the research.

    The reliance by Carey et al. on more general search terms (e.g. 'systems science', 'systems thinking', 'complex systems') without also including specific methodological names ('agent-based modeling', 'network analysis', 'community-based operations research') may have led to the high proportion (more than half) of more general introductory, expository, commentary-type publications included in the review. Therefore, the proportion of papers that engage in the rhetoric of systems science is likely over-stated and we believe there is a greater proportion of applications of systems methods than suggested by the review.

    Nonetheless, the large number of position papers is an important observation and we agree that public health has yet to take full advantage of systems science. Future studies are needed to examine the role of rhetoric in the public health research system and, for example, how these types of articles contribute to whether, to what extent, and in what contexts ideas such as systems thinking diffuse in public health. It is likely that discussion of methods new to a research field raises awareness, and possibly serves other functions (9). For example, discussion might influence selection, reformulation, and adaptation of methods from other fields.

    We commend the effort of Carey et al. to examine the state of systems science research in public health and agree with the authors that the review highlights important questions and future research directions. However, we caution that the findings in the recent systematic review likely present an unrepresentative view of the methods that comprise a systems science approach, important constituent approaches, and the relative prevalence of various approaches in the field. We encourage interested readers to search more extensively to find research and discussions of systems science methods. This is a difficult task, in part because of the diversity and evolution of terms and phrases authors use when writing about systems science. This situation could be remedied to an extent if terms indicating frequently used or emerging systems science methods were added to MeSH, as has been done for other methods such as clinical trials, meta-analysis, and community based participatory research. We also encourage practitioners of systems science methods to identify as such. This does not always happen, as emerging methods often need to be placed as more similar to commonly accepted scientific approaches, which are more likely to be deemed publishable. These changes may mitigate concerns that future reviews misrepresent the state of the science, and improve the ability of readers to access studies on the broad range of systems science applications in public health as language around systems science matures and new methods are developed.


    1. Carey G, Malbon E, Carey N, et al. Systems science and systems thinking for public health: a systematic review of the field. BMJ Open, 2015;5:e009002.

    2. Lich KH, Ginexi EM, Osgood ND, et al. A call to address complexity in prevention science research. Prev Sci, 2013;14:279-89.

    3. Luke DA, Stamatakis KA. Systems science methods in public health: Dynamics, networks, and agents. Annu Rev Public Health, 2012;33:357-76.

    4. Christakis N, Fowler J. The spread of obesity in a large social network over 32 years. N Engl J Med 2007;357(4):370-9.

    5. Orr MG, et al. Reducing racial disparities in obesity: simulating the effects of improved education and social network influence on diet behavior. Ann Epidemiol, 2014;24(8):563-9.

    6. Williams B, Imam I. Systems concepts in evaluation: An expert anthology. EdgePress, 2007

    7. Frerichs LF, Lich KH, Dave G, Corbie-Smith G. Integrating systems science and community-based participatory research to achieve health equity. Am J Public Health, 2016;106: 215-222.

    8. Speybroeck N, van Malderen C, Harper S, et al. Simulation models for socioeconomic inequalities in health: a systematic review. Int J Environ Res Public Health, 2013; 10(11):5750-80.

    9. Hyland, K. Disciplinary discourses: Social interactions in academic writing. Univ of Michigan Press, 2004

    Conflict of Interest:

    None declared

    Read all letters published for this article

    Submit response
  8. Some questions about the experimental design

    Will you be using Bayesian statistics for this study? If not, will you be publishing the confidence levels you'll be choosing, with a motivation as to why, before the study starts?

    It seems particularly important to have a picture of what you expect to find before the study starts. Over what period, for example, are you expecting any effects to occur? Are you prepared for effects to follow non -linear patterns - somebody may improve three months after surgery, but relapse to previous levels after eight?

    How will you cater for the baseline effect? Evidence for different levels of cognitive reserve suggest that the negative effect of obesity (if there is one) will be less apparent with those with high cognitive functioning, and much more apparent with those of low cognitive functioning. So you'd expect the recovery, on removal of any negative effect to differ between the extremes. A Bayesian approach would consider this, carefully, using previous evidence, so that an apparently large recovery in one group is seen in relation to the cognitive reserve, not simply as an absolute recovery.

    Prima facie, you'd expect those with very low initial cognitive ability, and thus very low cognitive reserve, to be most strongly affected by anything that affects cognitive ability. Their quick recovery, from that low baseline, if not properly handled, might give a false view of the chance of cognitive recovery being noticeable in those with large cognitive reserves - who might show very little effect.

    Are those control groups adequate? Presumably, it's important to exclude the effect of surgery itself, so a control group of individuals from the same population, who have similar surgery, but not specifically bariatric surgery, would help exclude that.

    Alternatively, to exclude the effect of surgery, would it make sense to have a control group that loses weight medically, rather than surgically?

    If you have a negative result. For example, it might be that the limitations on diet after bariatric surgery make the mood disorders worse, rather than better. Will you be reporting that?

    Will all the raw data be made available afterwards?

    How will you blind the structured clinical interviews? It seems that it would be quite easy for interviewers to notice whether those they interview are obese, or not. This is likely to result in a bias in the results, if the interviewers know anything about the nature of the study. Even if not, preconceptions of the link between obesity and depression might skew the results of the interviews.

    Have you a policy for dealing with any outliers? As discussed above, those with very low, or very high, cognitive reserve are likely to end up at the extremes. However, if you take the cognitive reserve into account, they may be useful and reasonable data. If the policy is to discard, say, 1% of outliers, then this information will be lost. Will the policy for dealing with outliers be published in advance of the start of the study?

    Conflict of Interest:

    No monetary interests. I do have the opinion that, if possible, medical solutions to obesity are to be preferred to surgical ones.

    Read all letters published for this article

    Submit response
  9. Response to: Moxey JM, Jones LL, 2016. A qualitative study exploring how Somali women exposed to female genital mutilation experience and perceive antenatal and intrapartum care in England.

    We welcome this important study into how Somali women affected by female genital mutilation (FGM) experienced accessing antenatal and intrapartum care within their area. The convenience and snowball sampling methods used to recruit in this study are both recognised, and appropriate, methods of sampling a population that is difficult to access. However, these methods have limitations which were not made explicit, for example, using a community centre leader within the social circle. We appreciate that this initial approach was important to gain the confidence of the local community and establish rapport, however, recruitment may have been improved by extending to other communities across England, as suggested by the title, or the United Kingdom.

    This study also aimed to identify factors which barred this population from accessing such care services - however, we feel that this aim was separate to the former objective, and therefore, in terms of the methodologies chosen, we consider that a mixed methods study design would have been more appropriate. While semi-structured interviews are suitable for maintaining confidentiality, and reducing the influence of cultural pressures on findings, the use of focus groups would have been a better method to explore factors affecting access to antenatal care services. Equally, on reading the study findings, it is noted that there is limited focus on the experiences of participants in antenatal care, for this reason, perhaps a separate study exploring these factors is warranted. However, with 10% of maternal deaths reportedly occurring in the non- English speaking migrant population, and 35% of these being preventable if antenatal care had been accessed earlier (1), such a study is an appropriate 'springboard' for further research, as very few previous studies exist that explore this area. Such research could also involve comparative studies to identify discrepancies in cultural perceptions of FGM within this population, such as between genders and generations.

    Additionally, no indication was given of the number of people approached and the number who declined to participate. We consider that collecting additional demographic information about the participants, such as sub-group of the Somalian community they came from, age of arrival to the UK and educational status would have provided context for interpretation of the findings although we appreciate that such profiling could affect confidentiality and sensitivity to the cultural norms of a small community.

    We questioned whether an interpreter was truly needed, being present at five out of seven interviews, but only translating in three. Moreover, it was pointed out that the interpreter was 'trusted by the community', further clarification on how this might have influenced findings might have added to data interpretation. While the researchers assessed their own positionality in relation to the participants, and how this impacted participant response, more discussion regarding interpreter positionality would have been useful. It would also have been important to consider, and perhaps justify, how the presence of a stranger or healthcare worker as interpreter may have influenced interview outcomes, such as building rapport or negative perceptions of healthcare staff and the legalities associated with FGM. However, we agree that adjustment of the topic guide following initial responses, and the use of double coding in data analysis may have countered this to a certain extent.

    Ultimately, the authors have provided readers with a clear statement of findings, but as previously noted, further contextual information, such as the grade of FGM, and attitudes to laws surrounding FGM in the UK, might have been instrumental in explaining any divergent findings. In terms of public health priorities, this raises issues of how to reach vulnerable groups, in order to build rapport and raise awareness within these communities. It is equally important to foster awareness on a local and national level - by working with organisations committed to combating FGM. Awareness of deinfibulation services and recommended guidelines (2) also needs to be raised, as women frequently present too late for this service - thereby increasing health risks and intrapartum complications. Ultimately, the priority for future service provision is integrated training and liaison between community, healthcare and other services to maximise identification and management of those at risk.


    1. Lewis G. The Confidential Enquiry into Maternal and Child Health (CEMACH). Saving Mothers' Lives: reviewing maternal deaths to make motherhood safer 2003-2005, the seventh report on confidential enquiries into maternal deaths in the United Kingdom. 2007.'%20Lives%202003 -05%20.pdf (Accessed online: 30th January 2016)

    2. Royal College of Obstetricians and Gynaecologists. Female Genital Mutilation and it's Management, Green-Top Guideline No. 53. 2015. (Accessed online: 21st January 2016)

    Conflict of Interest:

    None declared

    Read all letters published for this article

    Submit response
  10. Re: Health Checks

    We thank Professors Holland, Bewley and Capewell for their comments (2/2/2016) on our paper.(1) They consider that it is "...unfortunate that the study relies on routine general practice data collected by an electronic database. Although supposedly validated, there is no evidence of any check on the variability of recording by different practitioners or the validity of the Read codes used."

    From William Farr to Geoffrey Rose, the 'dirty hands' of epidemiologists have made good use of routine data like this. The QRISK2 algorithm, at the core of the Health Check, is also based on this routine data, validated by two groups of independent researchers who concluded that it more accurately predicted outcomes in diverse UK populations than the experimentally controlled Framingham dataset based on a single small American town. With the exception of family history of ischaemic heart disease, recording of major CVD risk factors at the Health Check was over 90%.

    Prof Holland and colleagues refer to "...some strange anomalies and classify new diagnoses in those undergoing a Health Check as an "outcome" (table 4). For example, although more individuals were classified as having type 2 diabetes in the Health Check group more individuals who had no Health Check had abnormal blood glucose levels".

    In an observational study, the measurement of risk factors by GPs in non-attendees is highly selective as they are based on symptoms or visible signs (obesity for example) which lead to high levels of diagnosis in a relatively small proportion of the non-attendees. Systematic population testing in the NHS Health Check therefore yields different results from testing in patient initiated consultations among non-attenders.

    In the 214,295 people attending the Health Check, 100,240 (46.8%) had a fasting or random blood glucose measurement at or within one year of the Check, compared to only 181,520/1,464,729 (12.4%) of the non-attendees. 1375/100,240 (1.4%) of attendees were found to have an abnormal test compared to 4274/181,520 (2.4%) of non-attendees. In other words almost four times as many attendees were tested but had a lower test positive rate than non-attendees. This means that the non-attendees tested were a relatively smaller and highly selected group, more likely to be positive than attendees. Even though the test positive rate was lower in attendees, many more were tested which resulted in twice as many new cases of diabetes within one year of the Check in the attendees 1934/214295 (0.9%) than non-attendees 5647/1464729 (0.4%).

    N. Ismail was concerned about apparently missing individuals in the flow chart. This occurred because of rounding. The data are given without rounding below.

    Total registered patients aged 40-74 years 2,411,030

    Total registered patients prior with NHS Health Check 27,442

    Total aged 40-74 years without prior check 2,383,588

    Total registered patients with exclusion criteria 704,564

    Total eligible patients 1,679,024

    Total patients with NHS Health Check 214,295

    Total patients without an NHS Health Check 1,464,729

    Prof Holland and colleagues repeat the claim by Krogsboll et al. that negative 'health check' trials demonstrate that the NHS Check "...waste(s) a vast amount of resource...". They also cite the later Inter99 study as evidence of ineffectiveness of the NHS Health Check programme.(2,3) Most trials cited by Krogsboll were over 20 years old and conducted before statins were invented or modern antihypertensive drugs routinely used. Of the trials since 1994, only one specifically recommended drug treatment for CVD risk, the other three offering advice but no drug treatment.

    The Inter99 trial published in 2014, illustrates the difficulty of mounting reliable trials of behaviour modification in populations at low risk to demonstrate reductions in major CVD endpoints. Unlike the NHS Health Check, Inter99 included people already on treatment for heart disease, stroke, diabetes, hypertension, heart failure and those on statins. The intervention was counselling with no specific recommendations for treatment. Most of those in the intervention group, 65%, did not receive the counselling intervention as planned or at all.(4) Those who did attend counselling showed twice the sustained validated quit rate in smokers and important improvements in alcohol consumption, diet and physical activity, all without increasing social inequity.(5,6)

    Inter99 demonstrated a 37% reduction in total mortality among attendees compared to controls which remained after adjusting for all known confounders.(7,8) When analysed to compare the two total randomised groups - all invited (irrespective of attendance) versus control - no CVD benefit was apparent.(3) The authors noted that a beneficial impact on cardiovascular outcomes could not be excluded by this study.

    Prof Capewell has argued for optimising "High-risk strategies which are medically based and effective for persons with high CVD risk" to complement wider population health measures.(9,10) Indeed he has shown that in England, one third of the recent reduction (to 2011) in cholesterol, had been due to statins.(11) The NHS Check programme has contributed to further substantial increases in statin use since then.(12)

    Stamler, reflecting on the MRFIT study wrote "The crucial scientific findings to end the CHD epidemic are now available. The challenge and task is to apply them in all appropriate patient contacts and across all population strata to extend the progress to date."(13)

    For people at higher CVD risk,the NHS Health Check is a cost- effective means to deliver treatment and identify more co-morbidity earlier. For people at lower CVD risk the most efficient means to deliver effective care remains to be determined.

    'Doing nothing' is not the alternative; it simply cedes the territory to the purveyors of vitamins, quack diets and private check-ups - less effective or less equitable alternatives. The NHS Check has proved to be a feasible framework, capable of equitably delivering coherent health messages around smoking, diet, physical activity and alcohol and engaging support from trusted sources. We can all agree we need more research to evaluate the strengths and weakness of the NHS Health Check programme to achieve its aims more efficiently.


    1. Robson J, Dostal I, Sheikh A, Eldridge S, Madurasinghe V, Griffiths C, et al. The NHS Health Check in England: an evaluation of the first 4 years. BMJ Open 2016;6:e008840.

    2. Krogsboll LT, Jorgensen KJ, Gronhoj Larsen C, Gotzsche PC. General health checks in adults for reducing morbidity and mortality from disease: Cochrane systematic review and meta-analysis. BMJ 2012;345:e7191.

    3. Jorgensen T, Jacobsen RK, Toft U, Aadahl M, Glumer C, Pisinger C. Effect of screening and lifestyle counselling on incidence of ischaemic heart disease in general population: Inter99 randomised trial. BMJ 2014;348:g3617.

    4. Jorgensen T, Borch-Johnsen K, Thomsen TF, Ibsen H, Glumer C, Pisinger C. A randomized non-pharmacological intervention study for prevention of ischaemic heart disease: baseline results Inter99. Eur J Cardiovasc Prev Rehabil 2003;10:377-86.

    5. Baumann S, Toft U, Aadahl M, Jorgensen T, Pisinger C. The long- term effect of a population-based life-style intervention on smoking and alcohol consumption. The Inter99 Study--a randomized controlled trial. Addiction 2015;110:1853-60.

    6. Baumann S, Toft U, Aadahl M, Jorgensen T, Pisinger C. The long- term effect of screening and lifestyle counseling on changes in physical activity and diet: the Inter99 Study - a randomized controlled trial. Int J Behav Nutr Phys Act 2015;12:33.

    7. Bender AM, Jorgensen T, Pisinger C. Is self-selection the main driver of positive interpretations of general health checks? The Inter99 randomized trial. Prev Med 2015;81:42-48.

    8. Bender AM, Jorgensen T, Pisinger C. Effects of general health checks differ under two different analyses perspectives-the Inter99 randomized study. J Clin Epidemiol 2015.

    9. Capewell S, O'Flaherty M. What explains declining coronary mortality? Lessons and warnings. Heart 2008;94:1105-8.

    10. Capewell S, Lloyd-Jones DM. Optimal cardiovascular prevention strategies for the 21st century. JAMA 2010;304:2057-8.

    11. Kypridemos C, Bandosz P, Hickey GL, Guzman-Castillo M, Allen K, Buchan I, et al. Quantifying the contribution of statins to the decline in population mean cholesterol by socioeconomic group in England 1991 - 2012: a modelling study. PLoS One 2015;10:e0123112.

    12. Robson J, Dostal I, Madurasinghe V, Sheikh A, Hull S, Boomla K, et al. The NHS Health Check programme: implementation in east London 2009- 2011. BMJ Open 2015;5:e007578.

    13. Stamler J, Neaton JD. The Multiple Risk Factor Intervention Trial (MRFIT)--importance then and now. JAMA 2008;300:1343-5.

    Conflict of Interest:

    None declared

    Read all letters published for this article

    Submit response

Don't forget to sign up for content alerts to receive selected information relevant to your specialty interests and be the first to know when the latest research is published.