Dear Editor, We would like to comment several limitations of the work of Gomez-Perreta. The work is merely a re-analysis of the data already published in 2001 (Navarro EA 2001) (http://www.emrpolicy.org/science/research/docs/navarro_ebm_2003.pdf). The data was obtained in a small city in Spain, called La ?ora, located in Murcia with an estimated population of 20.000 people. Both publications have from the same bias. The data used was obtained after a survey carried out from the data of only 88 responders out 215 questionnaires, which only represents 0,42% of the total population of La ?ora. The main issue is that the population asked to participate in the study was not blinded to the objectives to the study, with the high risk of bias due to the participation.

Authors only took into account microwave exposure at home, specifically measured in the bedroom; but neither microwave exposure at work nor at sites of leisure time, was considered as additional and eventually more relevant potential sources of microwave exposures.

The authors point out some possible confounders, in particular persons who believed in the damage of exposure (reached by phone in February 2012), the use of the computer and the use of the mobile. However, the authors have not taken into account many other relevant possible confounders. Among them, personal circumstances (life events) that may affect their physical/psychological situations, causing also anxiety, depression, irritability, sleep disorders, etc., such as recent unemployment or loss of a closed-relatives, previous health problems and current pharmacological treatment. Moreover, the variables analyzed in the study (anxiety, troubles of sleep, irritability) are closely related between them. In addition to the above high risk of bias, the authors have used the called Santini survey (http://www.ncbi.nlm.nih.gov/pubmed/12168254), a non -validated questionnaire that was just been translated into Spanish. Moreover the analysis was based on isolated variables included in the survey but not on a specific value assigned to the so-called microwave syndrome, according to Santini survey. We unsuccessfully tried to find in Internet any kind of validation of the mentioned questionnaire. It would be useful to get any further information from the authors in this respect.

Finally, prevalence is most appropriate than incidence to describe the results in this study.

In summary, due to the limitations in the design of this study it cannot be fixed with advanced statistical test. We are of the opinion that the conclusions according to the authors achieved by the study are not such conclusive ones. The study merely proposes a hypothesis for future research works based on well-designed studies, in particular cohort studies.

Conflict of Interest:

None declared

Morgan and colleagues elegantly propose the glitch methodology as an effective means to identify process deviations that could impact on the quality or speed of surgical procedures (1). We support the authors' suggestion that the analysis of glitch patterns may facilitate the development of interventions to improve patient safety in the operating theatre across a range of domains.

The development of a common method indeed seems possible. Using a similar methodology, exclusively in vascular surgery, we have shown remarkably similar results (2,3). The temporal pattern of glitches reported in vascular procedures (40% within first quarter, 75% within first half), reflects our previous data, in which many early intra operative errors can be attributed to poor preoperative planning- an area that we are currently assessing through the use of a novel software application (PREPARE (4)).

Morgan et al.'s laudable efforts in this arena requires two independent observers to record intraoperative glitches, one with a background in human factors (HF) and one with surgical experience, with good agreement between them. Interestingly, the authors identify that the clinical observer consistently reports more glitches than the HF observer. We feel that experienced health professionals working 'at the sharp end' are ideally placed to identify problems with patient safety. For that reason, we developed a method of identifying intraoperative error that can be used by operating teams, as well as observers. The Imperial College Error CAPture record (ICECAP) prompts structured reflection of intraoperative error during post-operative debriefing sessions with team members (5), and is being used in our UK multi-centre patient safety study in vascular surgery (4,6). Similar to the glitch classifications used by Morgan et al., ICECAP consists of six primary categories of error: equipment; communication; procedure-independent pressures; technical; safety awareness; patient-related, with twenty sub-categories used as prompts. Operating teams have been able to recall 69.9% of all errors recorded prospectively by a trained observer and nearly 79% of major errors (5).

Team self-report of error has several benefits over prospective observation, not least the fact that teams can feedback in real time, but are also uniquely placed to understand the context of errors that occur which is important in determining their significance. We believe that our self-report methodology also helps to foster a culture of openness and transparency to support patient safety- a key recommendation of the recent Francis Report on Mid Staffs (7). Indeed the effect of culture may, in agreement with Morgan et al., explain differences in glitch rates between hospitals identified in their study.

In the light of our recent systematic review on surgery and technology operating failures (8), we agree with Morgan et al. that differences in glitch rates between sites and specialties are likely to be dependent on differences in personnel, procedures and equipment. Interestingly, Morgan et al. found that glitch rates varied significantly in vascular. We have identified significant differences in error rates between procedural sub-types in vascular (open versus endovascular) (3); the types of treatment the authors observed may have influenced their findings.

As Morgan and colleagues state, it is unlikely that we can improve patient safety by addressing only one category of glitch, but characterising error patterns and their significance are crucial first steps in developing the wider surgical safety agenda.

1 Morgan L, et al. Capturing intraoperative process deviations using a direct observational approach: the glitch method, BMJ Open 2013; 3: e003519

2 Morbi AHM, et al. Reducing Error and Improving Efficiency during Vascular Interventional Radiology: Implementation of a Preprocedural Team Rehearsal. Radiology. 2012:264(2):473-483.

3 Albayati MA, et al. Identification of patient safety improvement targets in successful vascular and endovascular procedures: analysis of 251 hours of complex arterial surgery. Eur J Vasc Endovasc Surg. 2011;41(6):795-802.

4 Cardiovascular Safety Group Research Programme. Accessed Jan 16, 2014. http://www1.imperial.ac.uk/medicine/about/institutes/patientsafetyservicequality/research_themes_2/clinical_programmes/surgery/cardiovascular_safety_group/

5 Mason SL, et al. Design and validation of an error capture tool for quality evaluation in the vascular and endovascular surgical theatre. Eur J Vasc Endovasc Surg. 2012; 45(3):248-254.

6 UK Clinical Research Network. UK LEAP. Accessed Jan 16, 2014. http://public.ukcrn.org.uk/search/StudyDetail.aspx?StudyID=12982.

7 Francis R, The Mid Staffordshire NHS Foundation Trust Inquiry, 2013. Accessed Jan 16, 2014. http://www.midstaffspublicinquiry.com/report

8 Weerakkody RA, et al. Surgical technology and operating-room safety failures: a systematic review of quantitative studies. BMJ Quality and Safety 2013; 22(9): 710-718.

Conflict of Interest:

None declared