Appraisal of Evidence Skin tolerance and effectiveness of two hand decontamination procedures in everyday hospital use
Appraisal of Evidence
‘Skin tolerance and effectiveness of two hand decontamination procedures in everyday hospital use’
M. Winnefeld, M.A. Richard, M. Drancourt, J.J. Grob 2000
British Journal of Dermatology 143 546-550
For this assignment I am going to appraise an article from the British Journal of Dermatology, 2000, 143, 546-550. To locate this piece of research it was important to adhere to a focused search strategy. Due to the huge amount of research available today, in books, journals and online databases, it is simply not feasible to look through everything to discover a piece directly relevant to you (Flemming 1999). To retrieve information that is related, it helps to develop a focused question. The PICO framework devised by Sackett et al (1997) is useful in helping you to do this (cited by Craig and Smyth 2002). It helps you to focus on specifics, the population you want to look at, the intervention, a comparison intervention should there be one, and finally the outcomes of the research. For instance, you know you want to look into the area of infection control and this framework helps you to create a focus surrounding this. For example, the population you want to look at may be nurses and healthcare professionals, and their daily hand washing routine. A comparison may be introducing a new or alternative hand hygiene method. Finally the outcomes you would like to see from this research, it could be whether one method is more effective at disinfecting hands.
From a focused question you can choose key words. These can be used as search terms in data bases to retrieve pieces of specific research. Each data base contains some different pieces of information (Craig and Smyth 2002), so it is important to keep a list of the data bases and search terms used. You may only want resent research so specifying a year to search from, and also setting language limitations helps to facilitate the search. From the above focused question, I choose the search terms, ‘nursing staff, hand disinfection, everyday’, then entered these terms into MEDLINE. This retrieved the article under appraisal.
The aims of the study are clearly set out within the research, to asses the properties of an alcohol-based disinfectant, compared with a non-antiseptic soap at eliminating bacteria’ and, their effect on the skin. It is very important that research clearly states its aims so that you can assess whether these have been met by the end of the piece (Thompson 1999). The reasoning behind why the research was carried out is also discussed, with an explanation surrounding bacteria, and how it is carried and transferred in the hospital. It explains that damaged skin is a factor in higher levels of bacterial activity on hands, making skin tolerance important to optimum performance of hand disinfectants.
There were 78 people contacted to take part in the study. 52 people agreed to participate, only two of which were male. The participants came from a total of 16 different departments. However, as these were all in the same hospital, it could be suggested that this was a sample of convenience. Although sample size can not be relied upon by itself to give accurate findings, it is important that the sample gives an overview of the population it intends to study (Thompson 1999). In this case, 52 is a small sample and often large sample sizes are needed to get valid results from intervention studies (Thompson 1999). From the outset this is shedding doubt on the validity of any research findings. For a sample to represent the broader population, the sample must be selected using a random selection process (Cormack 2000). This would mean that everyone within the population would have had an equal chance of being involved in the study, thus eliminating any sampling bias. As there is no mention of how the sample was chosen for this piece of research, we can not eliminate the possibility of sampling bias. It would therefore be difficult to generalise any findings from this study to the broader population of nursing staff.
This is a preview of the whole essay
Another issue is that of volunteer bias, the possibility that non responders may not be as healthy as the participants themselves (Thompson 1999). In this instance 26 out of 72 people contacted did not take part. It may be that some of these people had sensitive skin or skin problems, making them less inclined to participate. Should this be the case and these people had taken part, it could have made a difference to the overall results regarding skin condition.
The design used to collect the data was a Randomised Control trial (RCT). This is a quantitative approach, allowing findings to be transferred into numerical data and statistics (Cormack 2000). This type of design is good to use when making a comparison (Deane 2006), making it applicable to this particular study, where the properties of two hand disinfecting agents are being assessed. This design is particularly helpful for the bacterial samples collected for this research, and thus converting these results into numerical data for comparison. RCT’s are often thought of as a good design for data collection, as they are useful in helping to eliminate conscious and unconscious bias (Nelson 1999), however, this can not be taken at face value because, as I highlighted earlier in the appraisal, there are some issues of sampling bias with this research. Although the RCT is an effective tool for analysing a clinical intervention, it is not without its limitations. Leach (2006) discuss’ how it is difficult to include personal experience and attitudes with this method. This is often due to a lack of lee way in regards to answering questions, and responses to scales etc. This could have been a factor for individuals and the assessor when using the 2 scales to measure skin damage. One scale had 0-4 levels and the other had 3 components each marked out of 7. Should the damage sit between levels one would have to choose the nearest option, where as a qualitative approach would have allowed for more intricate answers.
The study indicates that the 52 volunteer’s were randomised to either the alcohol-based antiseptic hand rinse, or the non-antiseptic liquid soap. This should mean that the participants were randomly assigned to either of the products. However, there is no mention of how they conducted this, so it is impossible to be certain this was done accurately and fairly. However they do make us aware that there is an even distribution of males, females and age ranges within each group, meaning that there were no significant differences between the groups prior to the intervention. Nelson (1999) explains that this is important because any risk factors should effectively be split between the two groups, however this is not inevitable with smaller numbers. This suggests that with a smaller group, like this study, there is less chance of risk factors being evenly spread. Also it would have been useful to have the age range of the participants to assess validity of results for generalisation.
The procedure each of the groups would need to follow when washing their hands is defined in the research, rubbing hands at least 10 times with the non-antiseptic soap, and 3-5ml of the alcohol-based rinse ensuring all areas are covered. The amount of times hands were washed was recorded in individual diaries by participants. This in itself could bring about inaccurate findings. If a nurse or nursing assistant is very busy, it may mean that the hand washing procedure was not completed to the specification, or there may sometimes have been insufficient time to record it. As stated at the beginning of the research paper, although it is mandatory for hands to be washed after every patient interaction, in reality this occurs less than 50% of the time. These factors make it difficult to rely on the diaries for correct information about every hand wash, if indeed they were all recorded.
Bacterial samples were taken from the participants hands on 2 occasions. Firstly, after being separated into their two groups, and at the end of the final day, these all followed the Larson bag technique. These were all conducted in the same manor, suggesting the results should be reasonably fair within the different groups. However, it does not allow for the possibility that volunteers may have had busier days within the 8 day trial, meaning some days hands may have carried a greater amount of bacteria and not been through the test procedure. A trial running for 8 days is relatively short, so if the bacterial samples had been obtained at the beginning and end of each day, the results may have been more varied. It could be suggested that they may have been more accurate and conclusive.
The Larson Scale was also used in conjunction with skin tolerance. This test has been used in similar studies to asses skin condition (Creedon 2005) suggesting that it is an appropriate tool to use. These assessments were also made on the first and last day, posing the same issues as the bacterial count, in that some days mid trial could have been worse than the days where the skin was assessed. If a nurse had an extremely busy day mid trial they could have washed their hands many more times than the day of the Larson Score test. Also the subjects themselves used this scale to evaluate their own skin condition, but this could never be wholly accurate as each individual opinion will differ, meaning some people may have a very similar skin condition but one person may rate theirs as a 4 whilst the other rates theirs as a 3. If they had been assessed on a daily basis more (or less) variation of the individuals’ skin condition may have been apparent.
The transepidermal water loss was measured using an Evaporimetre. This type of tool is often used when assessing the damage to premature babies’ skin (Nonato 2001), so it is effective in measuring skin damage through hydration, which is pivotal to this piece of research. The research states that the assessment took place before the first and last day of the trial so any physiological changes during the working day did not affect the outcome. A limitation of excluding these could be that the physiological changes are important factors for the nursing staff as individuals, and any discomfort that arises from the daily routine will impact on their home lives, thus affecting how willing they are to continue with the hand disinfecting regime. Also, it could be suggested that it is not a fair test if conducted before a shift as there is no mention of how volunteers were instructed to treat their skin at home. It may be that some used moisturisers etc before and after work, which would clearly impact on the skins hydration, in tern affecting the results. Bearing this in mind it could make the results from the evaporimetre inconclusive.
Ethics committees look into whether a piece of research will in any way diminish peoples’ human rights or place people under stress that could have been avoided (Cormack, D. 2000). Although it is hard to get research published if it has not been passed by an ethics committee, there is no mention within this piece stating that it was. Therefore it is not possible to tell whether it was carried out in light of this. This is also supported by the fact there is no mention of a cover letter being sent out to participants, making sure that they understood the aims of the study and how it was to be carried out. There may have been a cover letter sent out, but as it is not acknowledged in the study we can not assume there was. If this had been mentioned it would have made the reader aware that all volunteers entered into the study fully aware of the processes they were going to undertake.
There are quite a few things missing from this research that are important to its validity, for example there is no mention if a pilot study was undertaken. There may have been a pilot study but as there is no mention of this in the paper it is impossible to tell. Pilot studies are very useful in helping to highlight any problems that may occur, allowing these to be rectified before the research study commences (Cormack 2000). If there was a pilot study for this particular research it could have been very informative about the methods used, such as the diaries and how constructive these were towards the final results. This would have helped to prove how valid a tool these were, or show them as a weakness and an alternative method could have been sort prior to starting.
The results section of the research makes it clear that there were to be 52 participants in the study, however one was unable to complete the trial as they were not able to make work over the 8 day period, nothing related to the study. It is important that this information was given in the study as dropout rates can affect the final results, making treatments appear more effective (Nelson 1999).
In regards to amount hands were washed, the paper states that there was no significant difference between the groups. We are given the 2 means for both groups as follows: liquid non-antiseptic soap as 10.11 with a standard deviation (SD) of 4.47 and that the respective values for the alcohol-based rinse was 10.11 and 3.44. Here we can see that there is a small SD which can suggest that most the values (amount of times individuals washed their hands) are clustered around the mean (Barnard 2001). If this is the case then it suggests most people washed their hands a similar amount of times over the 8 day period. However, it must be taken into consideration that mean and the SD are affected by extreme values (Cormack 2000), which could suggest that one person in each group washed their hands substantially more than another member, who washed significantly less times than the mean. These values would still balance out when calculating the mean. Without looking at their personal diaries it is impossible to tell if this occurred, and if so, whether it had any impact on skin tolerance.
The results from the three scales used to asses the skins condition varied from each other. The Larson Score when assessed by the observer showed no real difference between the groups. The individual assessments using this scale showed the alcohol-based rinse to be better tolerated by skin. This links into what I highlighted earlier, that personal opinions vary, making it hard to draw anything conclusive from this test. The Sauermann score showed that skin conditioned worsened quite considerably with the liquid non-antiseptic soap, which tallies in with the individuals assessment using the Larson Score. The difference recorded between groups using TEWL was deemed insignificant, the discussion explains that this could be in relation to the test period being to short, which means that it is not possible for these results to be taken as conclusive. The Bacterial samples showed that the alcohol-based rinse was more effective at 31% decontamination than the non-antiseptic soap at 20%. This has a p value (the possibility that these results could happen by chance) of 0.016. This is very low, as it is said to be significant statistically if it is under 0.05 or 5% (Craig and Smyth 2002), however as the sample is so small it is not possible to assume that this makes the result clinically valid.
In conclusion, questions have been raised regarding sampling bias, volunteer bias, whether results are conclusive as well as certain ethical issues. The research mentions that the products under comparison are widely used, however both products are French, and the test was carried out in one hospital in France. There is no evidence within the research to suggest that these products are used or available in other countries. There is also no mention of how much these hand disinfectants cost, so it is not possible to see if they are cost effective, which is central to Evidence Based Practice (Craig and Smyth 2002). Although the researchers imply that the results show the alcohol-based rinse is the more effective method unless skin is substantially damaged, I believe that the results are not reliable enough to draw this conclusion. Collectively, the factors mentioned in this appraisal help to highlight why the results can not be generalised further than the initial sample.
Barnard, C. Gilbert, F. McGregor, P. (2001) Asking Questions in Biology: Key Skills for Practical Assessments and Project Work Second Edition, Pearson Education Limited: Harlow.
Cormack, D. (ed) (2000) The Research Process in Nursing Fourth Edition, Blackwell Science: Oxford.
Craig, J. V. Smyth, R. L. (eds) (2002) The Evidence-Based Practice Manual For Nurses. Churchill Livingstone: London.
Creedon, A. ‘Healthcare workers hand decontamination practices: compliance with recommended guidelines’ Journal of Advanced Nursing 51(3) 200-216 [online] Blackwell Synergy (02 January 2007).
Deane, K. H. O. ‘Randomised Controlled Trials: Part 1, Design’ British Journal of Occupational Therapy 69(5) 217-223.
Flemming, K. ‘Searchable questions’ Nursing Times Learning Curve 3(2) 6-7.
Leach, M. J. ‘Evidence-based practice: A framework for clinical practice and research design’ International Journal of Nursing Practice 12 248-251 [online] Blackwell Synergy (20 December 2006).
Nelson, E. A. ‘Randomised controlled trials: Questions for valid evidence’ Nursing Times Learning Curve 3(5) 6-8
Nonato, L. B. ‘Transepidermal water loss in the intensive care nursery: measuring techniques and recommendations’ New born infant nursing reviews 1(1) 11-20 [online] CINHAL (2 January 2007).
Thompson, C. ‘Questioning Evidence: its validity and importance’ Nursing Times Learning Curve 3(4) 4-6.
Thompson, C. ‘If you could just provide me with a sample: examining sampling in qualitative and quantitative research papers’ Evidence-Based Nursing 2(3) 68-70 [online] bmjjournals (15th December 2006).