Skip to content

Data Handling Homework Year 1985

Students' Motivations for Data Handling Choices and Behaviors: Their Explanations of Performance

Leslie Keiler* and Brian Woolnough

Karen Kalumuck, Monitoring Editor


Cries for increased accountability through additional assessment are heard throughout the educational arena. However, as demonstrated in this study, to make a valid assessment of teaching and learning effectiveness, educators must determine not only what students do, but also why they do it, as the latter significantly affects the former. This study describes and analyzes 14- to 16-year-old students' explanations for their choices and performances during science data handling tasks. The study draws heavily on case-study methods for the purpose of seeking an in-depth understanding of classroom processes in an English comprehensive school. During semistructured scheduled and impromptu interviews, students were asked to describe, explain, and justify the work they did with data during their science classes. These student explanations fall within six categories, labeled 1) implementing correct procedures, 2) following instructions, 3) earning marks, 4) doing what is easy, 5) acting automatically, and 6) working within limits. Each category is associated with distinct outcomes for learning and assessment, with some motivations resulting in inflated performances while others mean that learning was underrepresented. These findings illuminate the complexity of student academic choices and behaviors as mediated by an array of motivations, casting doubt on the current understanding of student performance.

Keywords: secondary, motivation, laboratory, performance, assessment


With the national and international cries for accountability in education, student performances at all levels and in all school contexts are coming under ever-closer scrutiny. The outcomes of student performances are subjected to detailed analyses and published in newspapers, becoming a factor in school selection and funding. The consequences for various levels of performance range in effect from student promotion and graduation, to teacher pay scale, to school accreditation. However, in the midst of all the discussion about what students do, there remains little understanding of why they do it. Even less do the impacts of various student motivations on learning and the accuracy of performances in representing learning factor into the political debate. The current study explores these issues in the context of secondary school science.

In the United States, the United Kingdom, and other educational systems around the world, “practical work” has come to play an increasingly significant role in a variety of subjects, especially in science. Having begun with Science—A Process Approach (American Association for the Advancement of Science [AAAS], 1969) in the United States and carried on in projects such as the Nuffield Science projects (Wellington, 1998; Donnelly and Jenkins, 2001) and the Assessment of Performance Unit (APU) (Driver et al., 1982; Archenhold et al., 1988) in Britain, the procedural science movement continues to influence curriculum and assessment in science education worldwide.

Science in the National Curriculum (Department for Education, 1995) of England and Wales includes content goals and targets in biology, chemistry, and physics, but its first section at every level is Experimental and Investigative Science. In this primary section lie the procedural goals of handling data, including making decisions about what and how much data to collect, analyzing data through calculations and graphs, interpreting data by seeking patterns, and evaluating data for reliability. This section is assessed through the students' submission of course work and forms 25% of their final marks for their General Certificate of Secondary Education (GCSE). Through the assessment process teachers, Boards of Examiners, and interested others can ascertain how well students performed on data handling tasks in this particular assessment context. However, a detailed exploration of this system reveals that it does not begin to address the issues of determining the full extent of students' knowledge and skills in data handling or their reasons for implementing or failing to implement their complete range of competencies. The research described here, a naturalist study of 14- to 16-year-old students doing practical work in an English school, analyzes students' explanations for their data handling in an effort to understand their performances on these various tasks.


According to Head (1985, p. 31), “Both the ability to perform a task and a willingness to do so are necessary for success, the latter often proves the more important.” Motivation can be conceived of as a will they or won't they phenomenon (Cannon and Simpson, 1985) or “his/her willingness to engage in the relevant learning activities” (Hofstein and Kempa, 1985, p. 222); however, it is the development of classifications of types of motivations that illuminates students' complex behaviors and performances. Hodson (1998b, p. 55) makes the point that various motives have different results for student learning and performance:

We should also bear in mind that when students are presented with a learning task they may perceive it in a way that is in marked contrast to the way in which the teacher saw it during the planning stage. Consequently, their actions may be somewhat different from those anticipated. Rather than attending to the rational appraisal of competing explanations in order to extend their understanding, for example, students may be actively engaged in any number of other pursuits, including: seeking teacher approval for compliant behavior; trying to look busy, thereby avoiding unwelcome teacher attention; ascertaining the 'right answer' (that is, the one that gains marks in tests); trying to maintain feelings of self-worth; attending to their 'classroom image'. These other agendas may lead students to adopt behaviors and make responses that are not helpful in bringing about better scientific understanding.

As Weiner (1984, p. 18) explains,

It is evident that many aspects of student activity are quite logical and rational: strategies are consciously employed to deal with threat and anxiety; goal expectations are consciously calculated; logical decisions are reached; information is sought and processed; self-insight may be attained; and so on. Conversely, many aspects of student behavior appear to be quite irrational: self-esteem is defended in unknown ways; expectations are biased; illogical decisions are reached; information is improperly utilized; and there is gross personal delusion.

Yet the seeming illogic of some student behavior can be laid at the feet of the observer: “If we cannot specify an individual's goals, we cannot judge what behavior will maximise the chances of achieving these goals and minimise the chances of avoiding undesirable outcomes” (Nicholls, 1984, p. 40). Thus, examining why students engage in various classroom behaviors involves the dissecting of multiple overt and covert goals and agendas.

Maehr (1983) describes four goal types that he believes are associated with school achievement: task involvement, ego involvement, social solidarity, and extrinsic rewards. According to Maehr, when students are pursuing task goals they are absorbed in the activity and seek competence in the task for the sheer pleasure of doing well. Ego goals, however, involve “doing better than some socially defined standard, especially a standard inherent in the performance of others” (Maehr, 1983, p 192); here doing better or best is the motivator. Maehr describes social solidarity goals as being directed toward pleasing others, while extrinsic reward involves motivation by acquisition of something such as a high mark or extra free time. Nicholls (1983) combines the latter two goals into a single extrinsic involvement motivation. These distinctions among motivations, their implications for interpreting classroom behaviors, and their interactions with concepts such as attribution have been explored and expanded in theory and research during the past two decades. Ames and Archer (1988, p. 260) collapse the various classification systems, claiming that “the conceptual relations among task, learning, and mastery goals and among ego, performance, and ability goals are convergent,” and thus in their work use only “mastery” and “performance goals”, respectively. However, this composite leaves out much of the richness and complexity of students' motivations in classroom settings, especially with reference to performance goals. Deci et al. (2001a) analyze “intrinsic motivations” and “extrinsic rewards” but create many internal classifications depending on the type of reward and context. Maehr's framework of four goal types illuminates the subtle but important differences among motivations that are focused away from the task, yet even this set does not incorporate the goal of limiting effort that appears in research on mental models (Norman, 1983).

An additional possible explanation for student behavior falls outside the literature of motivation but, nevertheless, is worthy of consideration: rule following. White (1988, p. 38) defines rules as “procedures, algorithms, which are applicable to classes of tasks.” Scardamalia and Bereiter (1983, p. 63) go as far as to say that “the normal processes of acquiring procedural knowledge or 'know-how' include observation, practice, and rule learning.” However, while they acknowledge that rule development and following are normal parts of learning, they discuss the drawbacks to such rules with regard to reading and writing:

Here, too, there is much redundancy, so that with practice students can develop efficient strategies that allow them to meet the routine demands of school reading and writing tasks with a minimum of effort. The result, however, is comprehension strategies that are insensitive to the distinctiveness and complexity of text information (Scardamalia and Bereiter, in press c), and writing strategies that are insensitive to distinctive requirements of different writing goals (Scardamalia and Bereiter, in press a). Rising above these routine “cognitive coping strategies,” as we call them, requires sustained effort directed towards one's own mental processes. (Scardamalia and Bereiter, 1983, p. 65)

Thus, when rule following, students are not truly task involved because the rules may not actually be appropriate to the task. Scardamalia and Bereiter (1983, p. 65) warn that there are some “skill areas in which 'practice makes perfect' is an untrustworthy slogan.”

Davis (1988, p. 97) discusses another concern within the area of rules:

Rule-following requires the existence of an established use or custom. Understanding a rule is to possess certain abilities. Wittgenstein distinguishes between following a rule, where an agent “knows that there is a rule, understands it, and intentionally moulds his actions to it” (1958, p. 155), and merely acting in accord with a rule, as a monkey might move chess pieces on a chess board in a way which happened to conform with the rules.

Thus, if this is correct, a crucial difference exists between students who understand rules and those who merely follow them. The danger to educators is that the two performances may appear identical, eliminating the possibility of easily discerning learning from conforming.

There is overlap among these various theories of motivation and explanations of student behavior. Thus, a compiled literature framework for description of student goal types and behavioral explanations in classroom activity consists of 1) task involvement, 2) ego involvement, 3) social solidarity, 4) extrinsic reward, 5) effort minimization, and 6) rule following. However, some significant areas of contradiction and variation in emphasis exist among these researchers and their findings. For example, current meta-analysis authors disagree about the negative effects of extrinsic rewards on intrinsic motivation (Cameron, 2001; Deci et al., 2001b). Some researchers have found that motivations can coexist, while others claim that the appearance of one replaces another (Seifert, 1996). Additionally, whether rule following enables or inhibits student learning remains a contested issue (Scardamalia and Bereiter, 1983; Davis, 1998). Therefore, a thorough understanding of student performances from within the current environment requires further study and analysis of their motivations.


This motivation study is part of a larger project exploring the data handling choices and behaviors of 14- to 16-year-old students engaged in science activities in an English comprehensive school (Keiler, 2000). The study relies on qualitative case study methods of observations and interviews, with the motivation findings coming from interviews with student participants. A single school was selected so that an in-depth description of the students' experiences could be developed (Schofield, 1990; Wolcott, 1994), considering that “selection on the basis of typicality provides the potential for a good ‘fit' with many other situations. Thick description provides the information necessary to make informed judgments about the degree and extent of that fit in particular cases of interest” (Schofield, 1990, p. 211). The study school was selected based on certain criteria that maximized its “typicality.” According to its own literature, it “is a medium sized fully comprehensive school for pupils between the ages of 11 to 16 serving [a town] and the surrounding area. It is the only Secondary School in [the town].” Being the only secondary school in town diminished the effects that school choice and parent selection might have on the student sampling frame. The school had students from a wide range of economic backgrounds, but very little ethnic or language diversity. According to the school's information on the Science department, “At Key Stage 4 all pupils follow the NEAB Modular Science Course. This allows pupils to gain a double award in Science. Pupils study modules which cover aspects of Biology, Chemistry and Physics.” These modules, taught by teachers qualified in the subject area, spread the three subject areas over the 2 years comprising Key Stage 4.

Key Stage 4 students are 14–16 years old or in Years 10 and 11 of school. Key Stage 4 Science in the National Curriculum (Department for Education, 1995) includes sections on Experimental and Investigative Science, Life Processes and Living Things, Materials and Their Properties, and Physical Processes. These areas can be roughly translated into experimental procedures, Biology, Chemistry, and Physics, respectively, with a little Earth Science mixed in. Key Stage 4 comprises the final 2 years of school when all students are required to take science, which culminates in the General Certificate of Secondary Education (GCSE) exam. The final GCSE mark combines course work, called assessed practicals, with the examination score. By investigating Key Stage 4 students, the study allows for the maximum possible learning while avoiding the self-selection of the next level of education. The classes observed were preparing for the higher tier of their Science GCSE exam. According to School Curriculum and Assessment Authority (SCAA) (1995, p. 74) regulations, the lower tier includes only selected parts of each of the numbered areas of the Key Stage 4 program of study, while the higher tier “must address all aspects of these sections of the programme of study.” Thus, sampling the students preparing for the lower tier would have restricted the range of knowledge and skills the students were expected to display in Science. Limiting the range of tasks in Science lessons might have artificially reduced the use of data handling by the students, of which this study sought the maximum. This sampling of higher tier curriculum demarcated the range of students participating in the study, and claims developed from the work might be confined in application to this population. However, approximately 90% of the students at the school were preparing for the upper tier examination and, thus, included in the sampling frame. Students were tracked into classes based on past performance in Science, and classes from high, medium, and low levels were included in the study.

Twelve units of work, selected by the teachers at the school as involving data handling, were observed and both impromptu in-class interviews and semistructured out-of-class interviews were conducted (Merriam, 1988; Millar et al., 1994; Kvale, 1996). Impromptu interviews occurred while students were engaged in data handling during lessons, usually following up a comment made by a student to his/her classmates. These interviews consisted of a question or two. According to Kvale (1996, p. 27), “Technically, the qualitative research interview is semistructured: It is neither an open conversation nor a highly structured questionnaire.” This method was chosen because information was being gathered about a specific topic, data handling, but the researcher wanted to remain responsive to relevant issues raised by the interviewees. According to Merriam (1988, p. 74),

In the semistructured interview, certain information is desired from all the respondents. These interviews are guided by a list of questions or issues to be explored, but neither the exact wording nor the order of the questions is determined ahead of time. This format allows the researcher to respond to the situation at hand, to the emerging worldview of the respondent, and to new idea topics.

However, while the semistructured interview was used for all events that the researcher labeled “interviews” to the participants, the in-class impromptu interviews and out-of-class discussions with teachers more closely followed Wolcott's (1995, p. 106) “casual or conversational interviewing.” These events were much less directed by the researcher and, in many cases, provided context and topics for the semistructured interviews. In providing so much control to the research subjects, semistructured interviews access information that the interviewer may not have known was available. However, this structure limits the effectiveness of quantifying responses and making cross-interview comparisons, as the subjects may not choose to address identical topics. This type of qualitative research generates a broad description of phenomena, not necessarily an accurate estimate of frequency.

Three groups of two to five students who worked together in each class were asked to participate in semistructured interviews outside of class time. The students were interviewed as soon after observed class periods as could be scheduled without interfering with their other responsibilities, usually during their break or lunch the following day. The purpose of the student interviews was to ascertain the students' thinking and decision-making during the data handling portion of the unit of work. The same groups of students were interviewed at two or three points during the unit of work, to check their progress with the work, confirm their explanations, and compare their plans to their accomplishments. The timing and number of interviews were determined by the instructional events in the unit. The students were treated as experts on their own actions and learning and asked to explain their choices and behaviors since “interviews are a useful means of gaining partial access to the child's knowledge and attitudes” (Palincsar and Brown, 1989, p. 23). Interview questions were designed to ascertain the sources of information and skills demonstrated by the students, the students' thought processes as they handled data, and their affective responses to these activities. Shulman (1986, p. 17) suggests, “To understand why learners respond (or fail to respond) as they do, ask not what they were taught, but what sense they rendered of what they were taught.” It was this sense of their own learning experiences about which the students were questioned. Finally, the students were asked to evaluate their performances during the classes and provide suggestions for improving their work; this served to demonstrate some of the differences between what students can and what they do accomplish, as the students pointed out discrepancies between what they knew and what they produced. Samples of the students' work were reviewed during classes and interviews. These interviews lasted between 20 and 40 min, with interviews later in the unit lasting longer than those following data collection. Additionally, lessons reviewing for GCSE examinations were observed and students were interviewed immediately following these examinations for approximately 0.5 h. The units of work observed were as follows.

  • 1 Year 11 Biology class—assessed practical: enzyme catalysis

     (3 lessons, 6 student interviews, 2 teacher interviews)

  • 1 Year 11 Physics class—assessed practical: springs

     (4 lessons, 10 student interviews, 1 teacher interview)

  • 1 Year 11 Chemistry class—investigation: rates of reaction

     (2 lessons, 2 student interviews, 1 teacher interview)

     (This unit was cut short due to the death of the teacher.)

  • 3 Year 11 Chemistry classes—assessed practical: rate of reaction

     (9 lessons, 11 student interviews, 1 teacher interview)

  • 2 Year 11 Biology classes—inheritance problems

     (4 lessons, 4 student interviews, 1 teacher interview)

  • 2 Year 10 Physics classes—assessed practical: electrical resistance

     (10 lessons, 6 student interviews, 2 teacher interviews)

  • 2 Year 10 Biology classes—assessed practical: osmosis

     (5 lessons, 5 student interviews, 1 teacher interview)

  • 2 Biology, 3 Chemistry, 2 Physics classes—GCSE review sessions

     (10 classes, 4 student interviews)

In two cases a class of students appeared in more than one unit. Thus, excluding the review sessions, the study included 10 classes of students, with 60 students being directly involved through interviews and/or work samples.

Student sampling decisions were made based on detailed information from the study site. For example, the physical arrangement of the room and the number of students in work groups partially determined how many students could usefully be observed in one lesson. As Cooper and McIntyre (1996, pp. 28–29) find, the ideal of interviewing all students about the lessons was impractical and impossible. They explain their alternative:

In order to minimize the potentially negative effects of failing to interview all pupils a sampling procedure was operated. This involved gathering data from the teachers about their perceptions of individual differences among members of the teaching group, through interviews and brief written comments. On the basis of these data it was possible to ensure that the pupils interviewed were broadly representative in terms of the salient differences among them as perceived by teachers.

Stake (1994, p. 244) suggests for within case sampling that the: “researcher notes attributes of interest…. discusses these characteristics with informants, gets recommendations,…. The choice is made, assuring variety but not necessarily representativeness, without strong arguments for typicality” (Stake 1994, p. 244) thus, prioritizing the opportunity to learn. As the study's focus was the students' data handling, the most important student feature for which some sort of representative sample was desirable was data handling performance. During student selection, science teachers were consulted in order to ensure the inclusion of highly skilled, middle range, and low performing students in the groups interviewed, with consideration for gender balance influencing the selection.

The data analyzed by developing, testing, and modifying assertions about the students' explanations through multiple readings of the student interview transcripts (Tobin and Fraser, 1987; Tobin and Gallagher, 1987; Anderson and Burns, 1989; Maykut and Morehouse, 1994; Millar et al., 1994; Cooper and McIntyre, 1996). The percentage of students who provided explanation in each category was calculated for each class and the entire sample. The conclusion discusses the outcomes for student learning and performance associated with the explanation categories. Marked papers and student examination marks were unavailable due to student and teacher confidentiality issues, the honoring of which was a condition of school access. However, even if the school had provided students' examination scores, no question-by-question analysis is conducted by the Examination Boards. Thus, it would be impossible to ascertain whether a high or low score in Science was due to data handling proficiency or the other 75% of the examination material. In this study judgments of learning were based on students' claims of understanding, demonstrations of understanding in interviews, and samples of work reviewed during interviews and classes. Claims are not made about student scores, but about the quality and quantity of learning that appeared to occur in these circumstances.


Both spontaneously and in response to questions students provided explanations for their choices and behaviors with regard to handling data. These student explanations fall within six categories, labeled 1) implementing correct procedures, 2) following instructions, 3) earning marks, 4) doing what is easy, 5) acting automatically, and 6) working within limits. These categories emerged from the data and use student language as closely as possible. The categories and their combination form the bases for the seven assertions about student motivations while handling data in science activities.

Some Students Claimed to Base Their Data Handling on “Implementing Correct Procedures”

The “implementing correct procedures” category consists of explanations students gave when they provided what they believed to be accepted criteria for their decisions about data handling. Seventy-seven percent of the students interviewed made at least one statement that fell into this category, including 100% of the students whose work was not currently being assessed. While none of the students explicitly said that they were “implementing correct procedures,” the explanations supporting this assertion demonstrated the students' beliefs that there are right and wrong ways to handle data and they were doing it the right way. In some cases, these explanations were based in accepted scientific practice; however, even when the students' scientific facts were erroneous, the motivation behind the explanation was a desire to follow what they believed were “correct procedures.” These explanations appeared especially frequently in their discussions about how much and what type of data to collect.

Researcher: Why did you choose six different concentrations (of chemical solutions)?

Winston: Because we wanted to get enough so that we could see a pattern developing in our results, and we thought that would be the right number.

Winston knew that patterns would be important for later data interpretation and was seeking the correct number to allow him to proceed accurately. Students also gave “correct procedures” explanations when they described how to analyze and interpret data. For example, in this group, John spoke as his two partners nodded in accord.

John: We're recording the voltage across the wire and the amps. And we do it five times and we average out the results.

Researcher: Why do you do it five times?

John: So we get an average of all the results, because one might be a freak result and where you got everything wrong or something. So you do it to see if you get all the same numbers.

John and his group wanted reliable data and believed that multiple trials and averaging would allow them to avoid a “freak result” or the effects of their getting “everything wrong.” This represented a widely held belief among these students that averaging was done to reduce the impact of “freak,” “stray,” or “anomalous” results. While this is not the statistical rationale for averaging data in experiments, the students in this study claimed that they were conducting multiple trials and averages for this reason. The fact that their justifications were unscientific does not lessen their legitimacy for the students, who expressed their belief in these procedures with deep conviction.

As part of their data analysis, student had to select the type of graph to include in their write-ups, sometimes attributing the decision to following a “correct procedure”:

Norman: It's not discrete data that you have. The gas syringe could have any quantity of gas in it, so rather than drawing a bar chart, it doesn't jump from thirty-six to thirty-seven in less than an instant, but it will go through thirty-six point one two three four. Because I think it's a line graph for continuous data.

Norman, unlike all others interviewed, correctly attributed his decision to the type of data he collected; he knew that line graphs were used for “continuous data” and that is what he believed he had. Therefore Norman followed a “correct procedure” and constructed a line graph. While Norman's reasoning stood alone in its scientific validity, other students did give explanations that demonstrated a concern for good practice, with the most common being the claim that line graphs showed patterns more clearly than other types of graphs.

Some Students Claimed to Base Their Data Handling on What They Have Been Told to Do

The “following instructions” category consists of explanations in which students claimed that the basis of their choice or behavior was doing what they were taught or told to do. Seventy-four percent of the students provided this justification at some point in their interviews. Comments such as “We aren't going to play dot-to-dot because that's what [the teacher] doesn't like” (George) typified this form of motivation. Students' attempting to follow rules passed on by their teachers also falls into this category.

Ruth: I think that's how it—there is a kind of rule that you have to use. I think that's how it is. I think time goes up the side and the variables come along the bottom, but it might be the other way around. I have to ask about that before I do it. We have been taught that but I've forgotten which way it goes [laughing].

Thus, the rule takes the form of instructions by the teacher, allowing the student to avoid making the decision for herself by asking the teacher to repeat the rule.

Some students appeared to be heavily reliant on teacher instructions, even when doing supposedly independent work.

Researcher: At what point did you decide that you were going to do averages?

Jane: She [the teacher] kind of told us.

Cathy: She wasn't supposed to and everyone kept saying “should we do the average” and then I don't know

Charlotte: I think people just figured it out.

In her interview the teacher claimed that the students had asked her whether they should average their results and that she asked them the question back, a claim that was supported by tape recordings from the lessons. However, at least Jane remembered the interaction as having been told to do the averaging; she considered herself to be doing what she was told. While the students knew that their teachers were restricted by assessment conditions, they still tried to ascertain what the teachers expected; Had the teachers been allowed, what instructions would they have been provided?

Sometimes the practices that the students claimed to have been taught were correct scientific procedures, e.g., using lines of best fit; sometimes the practices were erroneous, e.g., always putting time on the Y-axis. The identifying factor for this category is that the students believed that they were doing what the teacher wanted and expected. The critical difference between this and the previous category is the location the of the authority for the choice. In the “implementing correct procedures” category, the authority was in the method itself. Students claimed to be doing what was right because it was right; they indicated that they had made the choice themselves because it was, in some absolute way, the right choice to make. In the “following instructions” category the authority was with the teacher. In a sense, the fact that the teacher had told them to engage in a certain behavior absolved them of the responsibility of making the decision themselves.

Some Students Claimed to Base Data Handling on Earning Marks

The “earning marks” category consists of explanations in which students' descriptions indicated that their behavior was directed by what they must do to earn good marks on their assessed practical or exams. Seventy percent of students made claims related to earning marks at least once in their interviews. Some student conversations revealed this as the main reasons for all work in the 2 years of preparation for GCSE examinations.

Veronica: So basically everything you write down is just trying to gain you extra marks. That's the general reason for the investigations.

Rosie: That's the only reason we're doing it [laughing] is for the mark.

Researcher: Do you ever do experiments in class that aren't written up as assessed practicals?

Veronica: We used to in the first and third year, but now it's just either learning things for the final exam or doing assessed practicals. I think the teachers used the experiments when we were younger just to interest us and make us learn, but now they don't have time to do that.

Rosie: They have to count.

Veronica: If you're not doing theory, then you're doing an assessed practical.

Rosie: The last two years are just aimed at GCSEs mainly. Everything goes to a GCSE.

The marking system seemed so prevalent for these students that they perceived that all their work was directed toward earning marks: “Basically everything you write down is just trying to gain you extra marks.” When they mentioned learning, it was for examination purposes; they saw that the days when their interest in the material mattered were long past—“now they don't have time to do that.” These students' perceptions may have been reinforced by teacher comments similar to one during a Physics lesson, when the teacher admonished, “This is important. This is your GCSE practical.”

For some students, marks served as a motivation to do high-quality work. One student discussed the extra care she was putting into her write-up for her osmosis practical because it was being marked.

Valerie: With my write-up, I was trying to get everything in from diagrams and trying to explain what I was doing, because I'm trying to, I can't remember what my other grades were with my other investigations. But I'm trying to get better every time, trying to fit more in, so I can get a good grade, get a better grade for it.

Valerie was motivated to do good work because of the marks she could earn. She wanted to “get better every time” in order to “get a good grade”; she wanted her performance to improve.

Although motivations for earning marks appear to increase efforts by some students, for others the marking scheme acted as an upper limit on performance. The latter groups explained that they knew much more than they demonstrated on their assessed practicals because “You don't really need to, to get good marks” (Frank). Other students claimed that the desire to earn marks was so great that they would deliberately do work that they thought was erroneous if it was higher in the marking scheme: “Even if there's something that might be better, you still have to do the stuff on the syllabus, because otherwise you don't get the marks” (Jane). Thus, for some students, the “earning marks” motivation took priority over the “following correct procedures” option.

Some Students Claimed to Base Their Behaviors on “Doing What Is Easy”

Just over half the students in the study explained their decisions about data handling using the words “easy,” “easier,” and/or “easiest.” For some students, these reasons indicated a desire to minimize effort on their part; i.e., “easy” meant that they had to do less work or work that, for them, was at a lower level.

Researcher: Why did you choose to do a line graph for this one?

Joe: It was easiest [laughing].

Researcher: Okay.

Joe: A bar graph's really practically unlikely; I like line graphs.

This student and others like him consistently made choices that allowed them to put the least effort possible into their work. However, when some students used these terms to justify their choices, their explanations revealed a desire for elegance in the process of handling data. Although another group made the same decision, claiming that what they did was “easiest,” their explanation was very different from Joe's.

Harriet: Because you can see, because the shape of the line can show you easily the patterns.

Karen: It's the simplest to understand really. You just look at it and you can see exactly what happens.

Although they used the same terms as Joe, Harriet and Karen communicated a desire to understand their data using their graphs: “You just look at it and you can see exactly what happens.” They did not convey the impression that personal preference was an acceptable justification. They did indicate an awareness of the next step in their investigations, keeping in mind the overall purpose of graphing: “The shape of the line can show you easily the patterns.”

Similarly, a student claimed to be using lines of best fit and line graphs for the subsequent ease of interpretation.

Researcher: Why are smooth lines better?

Frank: It gives a clearer indication. It's easier to draw comparisons between two lines that are smooth than two lines that perhaps intersect and bubble.

And later,

Frank: With a line graph it tends to be easier to see, with the steepness; if it's a steep curve then the reaction is happening quickly and if it shallows out then the reaction will begin to slow down. It's harder to see that with a bar chart or whatever. It tends to be an easy type of graph to interpret.

While using the term “easier,” Frank described a “correct procedures” concern with identifying patterns as he justified his use of line graphs with lines of best fit, suggesting that he planned to “draw comparisons between two lines” and relate the rate of reaction to relative gradients. Karen, Harriet, and Frank used language that allowed their explanations to fall within the “easy” category, yet both groups indicated that, in this case, they were more concerned with gaining a quality product from their work than with reducing their efforts.

Some Students Claimed to Follow Data Handling Procedures “Automatically”

According to approximately one-fifth of the students in the study, various data handling procedures had become automatic; when they did an investigation they did not have to think about what to do with their collected data.

Julie: In Maths, in Science, we've just been encouraged to do graphs for years, and now it just comes naturally whenever you do an investigation.

Laura: Yeah, you don't think, “Oh we have to do.”

Geri: You have to do a graph.

Julie: You just do and so you do stuff like predictions hypothesis naturally, as well. So it's just a way of showing the results that you've got.

They had been trained to do procedures that they had repeated so often that, these students claimed, the implementation of the procedures had become subconscious. Julie believed of graphing that “now it just sort of comes naturally,” while Laura's “you don't think…you just do” clearly communicated the removal of the conscious aspects of the process.

Another group, discussing their resistance investigation, made claims of automatic practice.

Jane: People just do it naturally now.

Researcher: What do you mean “do it naturally”?

Jane: It's just something you have to do [laughing]. That's about it, with a graph or anything.

Charlotte: Yeah, usually you find the average of any results you have.

Jane: Make graphs with averages. You can't make a graph with every single result, because you'd have ten million graphs.

Cathy: And we've been told, when we do an experiment or something, we always have to do it five times, so we always have five results on one thing but less actual at the end results. So you just do it just automatically.

According to Jane, averaging is automatic for students doing investigations, “People just do it naturally now,” with Cathy supporting this claim, “You just do it just automatically.” While “naturally” and “automatically' may not technically mean the same thing, these students used them interchangeably. Both of these groups claimed not to think consciously about what procedures to include in their investigations. While the initial impetus for averaging and graphing may have been following teacher instructions, these students indicated that teachers no longer had to tell them to do graphs and other parts of their investigations.

Some Students Claimed That Their Behaviors Were Limited by Contextual Factors

Almost two-fifths of the students in the study discussed contextual factors, such as time and equipment, in their explanations of their data handling. According to these students, they would have behaved differently if they had not been working under particular conditions, especially time limitations.

Frank: The time limit that we get, two lessons, and it's not enough to really do four to six experiments. We have fifteen to do, which is.

Norman: We're supposed to do it on our own, as well, which is stupid because there's only fifteen gas syringes between a class of thirty.

Frank: We're meant to share, and it doesn't work.

Norman: There's always some people larking around [laughing].

Frank: If we're given the time, then we do it properly but, it's too much pressure. There's not enough equipment. There's always never enough acid or anything.

In addition to their discussion of “larking around,” some students did indicate that certain time constraints were of the students' own making. The same group who complained that they were not given enough time and equipment to collect their data later explained their poor performance as their own responsibility.

Frank: I probably could have done a lot better if I had really had the time. Because we tend to do it the night before it is due really [laughing]. They give you six weeks but the longer they give you…

Norman: It doesn't matter. We're going to do it the night before anyway.

Frank: They might as well give it to you tomorrow. It would be a better experiment that way; you'd actually remember it.

Nevertheless, students used these factors to account partially for their data handling choices and behaviors. These choices included fabricating data for their practicals, about which they were embarrassed but justified their behavior by describing what they considered to be unreasonable working conditions.

Many Students Described Multiple Motivations for Their Data Handling

While students occasionally gave a single explanation for their behaviors and choices, they frequently provided more than one reason in consecutive interviews, the same interview, or even the same response to a single question. Only two students provided a single explanation for all their decisions, which was earning high marks. In some cases one of the multiple motivations the students discussed seemed to take precedence, while in others no clear supremacy emerged. Some students appeared to include both the “right” reason and the “real” reason, parroting the “correct” explanations they had been given but mentioning their own underlying motivations. For example, while explaining their decisions about how many data to collect for their resistance investigation, one group included both “correct procedures” and “earning marks” reasons

Jane: Accuracy. Make sure you get it right. It can be just really fluky or something like that.

Charlotte: Just in case.

Cathy: It could be the wrong power or something, for some stupid reason.

Jane: And you wouldn't know, unless you…

Charlotte: And you have to do it three times. You have to repeat it three times.

Jane: To get the accuracy and points and stuff.

Charlotte: Averages.

Researcher: You just talked about a whole bunch of things, so can you tell me more about the points?

Jane: Oh yeah. It's just like for marking. If you've just done it once then you don't have a very reliable source. So you don't get as many marks as if you did it a whole load of times, got an average, and said why you think they weren't all the same and why they were the same and what you think about it.

Jane's initial response was that they made their decisions based on what would be the most accurate. Her group's “correct procedures” explanations included incorrect rationales of avoiding “fluky” results and experimenter error, but they were about accuracy nonetheless. Later, Jane elaborated on the marking scheme, but she indicated a belief that the higher marks depended on more accurate procedures. It was unclear whether her primary motivation was being accurate or earning marks by being accurate. These multilayered motives exhibit the complexity of the problem of understanding students' choices and behaviors.

Table ​1 lists the percentages of students who provided at least one motivation in each of the categories by unit of work and overall.

Table 1

Percentage of students who provided at least one motivation in each category by unit of work and overall

Discussion of Outcomes

The various explanations that students gave were associated with specific choices and behaviors, resulting in identifiable outcomes for both learning and performance. These outcomes were revealed through student explanations of their performances, including discussions of work samples. These outcomes are summarized below and related to findings in the literature.

When students claimed to be “implementing correct procedures,” they attempted to produce the highest-quality work they could. They focused on the task, rather than external factors. Thus, their performances could be expected to reflect their knowledge and skills of data handling accurately. When students provided “correct procedures” explanations, teachers could accurately assess what the students' understood and could correct their misperceptions. This was one of the two most common categories of student explanations. This “implementing correct procedures” set of explanations most closely matches Maehr's (1983) task orientation motivation and the positive outcomes associated with intrinsic motivation described by Deci et al. (2001a).

When students claimed to be “following instructions,” they did not make choices for themselves but depended on what they believed a past or present teacher had told them. Thus, they were able to avoid responsibility for their work. Further, they relied upon recollections of teachings rather than true understandings of procedures, making them vulnerable to memory failures. Additionally, these students sometimes were able to produce work that they did not understand, simply by following a set of rules. Student work associated with this explanation provides no real insight into student understanding of science, merely memory for instruction. This category composed the overwhelming motivation for a minority of students and appeared sporadically in other interviews. For some students, this category corresponds to aspects of Maehr's (1983) social solidarity motivation, as seeking teacher approval appeared to be part of the explanation. For others, it corresponds more closely to the literature regarding the negative aspects of rule following (Scardamalia and Bereiter, 1983; Larson, 1995) involving limited mental activity and lack of effort.

When “earning marks” motivated students, they considered learning to be secondary to performance. Several students explained that they saw the marking scheme standards as the upper limit of performance. They were not willing to expend time and energy that were not rewarded by marks, so they did not demonstrate their full range of knowledge and skills. Students also admitted that the emphasis of the school system on earning high marks justified their fabricating experimental data. Thus, when students were motivated by “earning marks,” their performances frequently misrepresented their scientific understandings. This explanation competed with “implementing correct procedures” for being the most prevalent in interviews and most powerful for the students. Many characteristics of the “earning marks” motivation coincide with Maehr's (1983) description of extrinsic rewards orientations and support Deci and co-workers' (2001b) conclusions about the negative effects of rewards on intrinsic motivation.

When students claimed to be “doing what is easy,” they acted in one of two ways. For one set of students “easy” meant that they avoided challenges, resulting in poor performances unrelated to actual knowledge and skill levels (Norman, 1983; Loughran and Derry, 1997). Other students used words such as “easy” when they sought elegant and useful procedures, producing work at their highest levels, matching characteristics of task-oriented students (Maehr, 1983). This category demonstrates that not only must educators listen to students' explanations of their work, but they must listen carefully if they want to appreciate true levels of understanding.

When students claimed to be “acting automatically,” they did not make conscious decisions about their choices and behaviors. Sometimes this led to their efficiently using tacit knowledge to perform data handling tasks. In other cases, students applied “automatic” behaviors inappropriately. In both instances, the choices, behaviors, and their products were unmonitored by the students. These students' explanations of their performances seem to fall outside Maehr's framework, even with the addition of an effort minimization motivation. Rather, they appear to relate to the literature of tacit knowledge (e.g., Polanyi, 1962; Woolnough, 1989; Claxton, 1997).

When students claimed to be “working within limits,” they did not perform at their full potential. They used contextual limitations as an excuse to avoid accountability for the quality and quantity of their work. Additionally, they created limitations for themselves, which could further protect them from exposing their actual levels of knowledge and skills. These explanations may belong to Maehr's ego orientation, in that blaming contextual factors allowed students to preserve their egos, or to effort minimization, as focusing on external constraints permitted them to reduce their efforts.

Thus, it seems that only students who were motivated by “implementing correct procedures,” and some of those who were “doing what is easy” and “acting automatically,” produced work for their assessed practicals that accurately reflected their data handling knowledge and skills. In some cases, such as when they were “following instructions,” students' final products exceeded their understandings. More commonly, however, the students' level of performance on their assessed practicals was far inferior to their potential, either because the marking scheme did not included relevant mastered techniques or because the students were able to shield themselves from responsibility for their choices and behaviors. These findings have strong implications for the reliability of conclusions drawn about levels of individual, school, and program performances when students' motivations are not fully understood. Further, this research suggests that, to maximize learning and accurately assess students' understanding, educators must resist the temptation to motivate students through extrinsic rewards, be judicious in their provision of specific instructions and standards for success, and foster a desire in students to perform their tasks completely and accurately.


  • Adey P. Does motivation style explain CASE differences? A reply to Leo and Galloway. Int J Sci Edu. 1996;18:51–53.
  • American Association for the Advancement of Science . The Psychological Bases of Science—A Process Approach. AAAS; Washington, DC: 1969.
  • Ames C., Archer J. Achievement goals in the classroom: Students' learning strategies and motivation processes. J Educ Psychol. 1988;80:260–267.
  • Anderson L., Burns R. Research in Classrooms: The Study of Teachers, Teaching and Instruction. Pergamon Press; Oxford: 1989.
  • Bell J., Donnelly J., Johnson S., Welford G. In: Assessment of Performance Unit: Science at Age 15: A Review of APU Survey Findings 1980–84. Archenhold F., editor. Her Majesty's Stationary Office; London: 1988.
  • Cameron J. Negative effects of reward on intrinsic motivation—A limited phenomenon: Comment on Deci, Koestner, and Ryan (2001) Rev Educ Res. 2001;71:29–42.
  • Cannon R., Simpson R. Relationships among attitude, motivation, and achievement of ability grouped, seventh-grade, life science students. Sci Educ. 1985;69:121–138.
  • Claxton G. Hare Brain, Tortoise Mind: Why Intelligence Increases When You Think Less. Fourth Estate; London: 1997.
  • Cooper P., McIntyre D. Open University Press; Buckingham, UK: 1996. Effective Teaching and Learning: Teachers' and Students' Perspectives.
  • Davis A. Transfer, abilities and rules. J Philos Educ. 1998;32:75–106.
  • Deci E. L., Koestner R., Ryan R. Extrinsic rewards and intrinsic motivation in education: Reconsidered once again. Rev Educ Res. 2001a;71:1–27.
  • Deci E. L., Ryan R., Koestner R. The pervasive negative effects of rewards on intrinsic motivation: Response to Cameron (2001) Rev Educ Res. 2001b;71:43–51.
  • Department for Education . Science in the National Curriculum. HMSO; London: 1995.
  • Donnelly J. F., Jenkins E. W. Science Education: Policy Professionalism and Change. Paul Chapman; London: 2001.
  • Driver R., Gott R., Johnson S., Worsley C., Wylie F. Department of Education and Science; London: 1982. APU: Science in Schools: Age 15: Report No. 1.
  • Head J. The Personal Response to Science. Cambridge University Press; Cambridge: 1985.
  • Hodson D. Is this really what scientists do? Seeking a more authentic science in and beyond the school laboratory. In: Wellington J., editor. Practical Work in School Science: Which Way Now? Routledge; London: 1998a.
  • Hodson D. Teaching and Learning Science: Towards a Personalized Approach. Open University Press; Buckingham, UK: 1998b.
  • Hofstein A., Kempa R. Motivating strategies in science education: Attempt at analysis. Eur J Sci Educ. 1985;7:221–229.
  • Keiler L. S. Factors affecting student data handling choices and behaviors in Key Stage 4 science. University of Oxford; Oxford: 2000. D.Phil. thesis.
  • Kvale S. Interviews: An Introduction to Qualitative Research Interviewing. Sage; London: 1996.
  • Larson J. Fatima's rules and other elements of and unintended chemistry curriculum. 1995. Presented at the Annual Meeting of the American Educational Research Association, San Francisco, Apr. 23.
  • Loughran J., Derry N. Researching teaching for understanding: The students' perspective. Int J Sci Edu. 1997;19:925–938.
  • Maehr M. L. On doing well in science: Why Johnny no longer excels; Why Sarah never did. In: Paris S. G., Olson G. M., Stevenson H. W., editors. Learning and Motivation in the classroom. Lawrence Erlbaum Associates; Hillsdale, NJ: 1983.
  • Maykut P., Morehouse R. Falmer Press; London: 1994. Beginning Qualitative Research: A Philosophic and Practical Guide.
  • Merriam S. Jossey–Bass; London: 1988. Case Study Research in Education: A Qualitative Approach.
  • Millar R., Lubben F., Gott R., Duggan S. Investigating in the school science laboratory: Conceptual and procedural knowledge and their influence on performance. Res Papers Educ Policy Pract. 1994;9:207–248.
  • Nicholls J. G. Conceptions of ability and achievement motivation. In: Paris S. G., Olson G. M., Stevenson H. W., editors. Learning and Motivation in the Classroom. Lawrence Erlbaum Associates; Hillsdale, NJ: 1983.
  • Nicholls J. Concepts of ability and achievement motivation. In: Ames R., Ames C., editors. Research on Motivation in Education: Student Motivation. Vol. 1 Academic Press; London: 1984.
  • Norman D. A. Some observations on mental models. In: Genter D., Stevens A. L., editors. Mental Model. Lawrence Erlbaum Associates; London: 1983.
  • Palincsar A., Brown B. Instruction for self-regulated reading. In: Resnick L., Klopfer L., editors. Toward the Thinking Curriculum: Current Cognitive Research. Association of Supervision and Curriculum Development; Alexandria, VA: 1989.
  • Polanyi M. Personal Knowledge: Towards a Post-critical Philosophy. Routledge and Kegan Paul; London: 1962.
  • Scardamalia M., Bereiter C. Child as coinvestigator: Helping children gain insight into their own mental processes. In: Paris S. G., Olson G. M., Stevenson H. W., editors. Learning and Motivation in the Classroom. Lawrence Erlbaum Associates; Hillsdale, NJ: 1983.
  • Schofield J. Increasing the generalizability of qualitative research. In: Eisner E., Peshkin A., editors. Qualitative Inquiry in Education: The Continuing Debate. Teachers College Press; New York: 1990.
  • School Curriculum Assessment Authority . GCSE Regulations and Criteria. SCAA; London: 1995.
  • Seifert T. The stability of goal orientations in grade five students: Comparison of two methodologies. Br J Educ Psychol. 1996;66:73–82.
  • Shulman L. Those who understand: Knowledge growth in teaching. Educ Res. 1986;15:4–21.
  • Stake R. Case studies. In: Denzin N., Lincoln Y., editors. Handbook of Qualitative Research. Sage; London: 1994.
  • Tobin K., Fraser B., editors. Exemplary Practice in Science and Mathematics Education. Key Centre for School Science and Mathematics, Curtin University of Technology; Perth: 1987.
  • Tobin K., Gallagher J. What happens in high school science classrooms. J Curric Stud. 1987;19:549–560.
  • Weiner B. Principles for a theory of student motivation and their application within an attributional framework. In: Ames R., Ames C., editors. Research on Motivation in Education: Student Motivation. Vol. 1 Academic Press; London: 1984.
  • Wellington J. Practical Work in School Science: Which Way Now? Routledge; London: 1998.
  • White R. T. Learning Science. Basil Blackwell; Oxford: 1988.
  • Wolcott H. Transforming Qualitative Data: Description, Analysis, and Interpretation. Sage; London: 1994.
  • Wolcott H. The Art of Fieldwork. Altamira Press; London: 1995.
  • Woolnough B. Towards a holistic view of processes in science education (or the whole is greater then the sum of its parts, and different) In: Wellington J., editor. Skills and Processes in Science Education: A Critical Analysis. Routledge; London: 1989.

Articles from Cell Biology Education are provided here courtesy of American Society for Cell Biology

Persons using assistive technology might not be able to fully access information in this file. For assistance, please send e-mail to: Type 508 Accommodation and the title of the report in the subject line of e-mail.

Recommendations of the Healthcare Infection Control Practices Advisory Committee and the HICPAC/SHEA/APIC/IDSA Hand Hygiene Task Force

Prepared by
John M. Boyce, M.D.1
Didier Pittet, M.D.2
Hospital of Saint Raphael
New Haven, Connecticut
2University of Geneva
Geneva, Switzerland

The material in this report originated in the National Center for Infectious Diseases, James M. Hughes, M.D., Director; and the Division of Healthcare Quality Promotion, Steve Solomon, M.D., Acting Director.

The Guideline for Hand Hygiene in Health-Care Settings provides health-care workers (HCWs) with a review of data regarding handwashing and hand antisepsis in health-care settings. In addition, it provides specific recommendations to promote improved hand-hygiene practices and reduce transmission of pathogenic microorganisms to patients and personnel in health-care settings. This report reviews studies published since the 1985 CDC guideline (Garner JS, Favero MS. CDC guideline for handwashing and hospital environmental control, 1985. Infect Control 1986;7:231--43) and the 1995 APIC guideline (Larson EL, APIC Guidelines Committee. APIC guideline for handwashing and hand antisepsis in health care settings. Am J Infect Control 1995;23:251--69) were issued and provides an in-depth review of hand-hygiene practices of HCWs, levels of adherence of personnel to recommended handwashing practices, and factors adversely affecting adherence. New studies of the in vivo efficacy of alcohol-based hand rubs and the low incidence of dermatitis associated with their use are reviewed. Recent studies demonstrating the value of multidisciplinary hand-hygiene promotion programs and the potential role of alcohol-based hand rubs in improving hand-hygiene practices are summarized. Recommendations concerning related issues (e.g., the use of surgical hand antiseptics, hand lotions or creams, and wearing of artificial fingernails) are also included.

Historical Perspective

For generations, handwashing with soap and water has been considered a measure of personal hygiene (1). The concept of cleansing hands with an antiseptic agent probably emerged in the early 19th century. As early as 1822, a French pharmacist demonstrated that solutions containing chlorides of lime or soda could eradicate the foul odors associated with human corpses and that such solutions could be used as disinfectants and antiseptics (2). In a paper published in 1825, this pharmacist stated that physicians and other persons attending patients with contagious diseases would benefit from moistening their hands with a liquid chloride solution (2).

In 1846, Ignaz Semmelweis observed that women whose babies were delivered by students and physicians in the First Clinic at the General Hospital of Vienna consistently had a higher mortality rate than those whose babies were delivered by midwives in the Second Clinic (3). He noted that physicians who went directly from the autopsy suite to the obstetrics ward had a disagreeable odor on their hands despite washing their hands with soap and water upon entering the obstetrics clinic. He postulated that the puerperal fever that affected so many parturient women was caused by "cadaverous particles" transmitted from the autopsy suite to the obstetrics ward via the hands of students and physicians. Perhaps because of the known deodorizing effect of chlorine compounds, as of May 1847, he insisted that students and physicians clean their hands with a chlorine solution between each patient in the clinic. The maternal mortality rate in the First Clinic subsequently dropped dramatically and remained low for years. This intervention by Semmelweis represents the first evidence indicating that cleansing heavily contaminated hands with an antiseptic agent between patient contacts may reduce health-care--associated transmission of contagious diseases more effectively than handwashing with plain soap and water.

In 1843, Oliver Wendell Holmes concluded independently that puerperal fever was spread by the hands of health personnel (1). Although he described measures that could be taken to limit its spread, his recommendations had little impact on obstetric practices at the time. However, as a result of the seminal studies by Semmelweis and Holmes, handwashing gradually became accepted as one of the most important measures for preventing transmission of pathogens in health-care facilities.

In 1961, the U. S. Public Health Service produced a training film that demonstrated handwashing techniques recommended for use by health-care workers (HCWs) (4). At the time, recommendations directed that personnel wash their hands with soap and water for 1--2 minutes before and after patient contact. Rinsing hands with an antiseptic agent was believed to be less effective than handwashing and was recommended only in emergencies or in areas where sinks were unavailable.

In 1975 and 1985, formal written guidelines on handwashing practices in hospitals were published by CDC (5,6). These guidelines recommended handwashing with non-antimicrobial soap between the majority of patient contacts and washing with antimicrobial soap before and after performing invasive procedures or caring for patients at high risk. Use of waterless antiseptic agents (e.g., alcohol-based solutions) was recommended only in situations where sinks were not available.

In 1988 and 1995, guidelines for handwashing and hand antisepsis were published by the Association for Professionals in Infection Control (APIC) (7,8). Recommended indications for handwashing were similar to those listed in the CDC guidelines. The 1995 APIC guideline included more detailed discussion of alcohol-based hand rubs and supported their use in more clinical settings than had been recommended in earlier guidelines. In 1995 and 1996, the Healthcare Infection Control Practices Advisory Committee (HICPAC) recommended that either antimicrobial soap or a waterless antiseptic agent be used for cleaning hands upon leaving the rooms of patients with multidrug-resistant pathogens (e.g., vancomycin-resistant enterococci [VRE] and methicillin-resistant Staphylococcus aureus [MRSA]) (9,10). These guidelines also provided recommendations for handwashing and hand antisepsis in other clinical settings, including routine patient care. Although the APIC and HICPAC guidelines have been adopted by the majority of hospitals, adherence of HCWs to recommended handwashing practices has remained low (11,12).

Recent developments in the field have stimulated a review of the scientific data regarding hand hygiene and the development of new guidelines designed to improve hand-hygiene practices in health-care facilities. This literature review and accompanying recommendations have been prepared by a Hand Hygiene Task Force, comprising representatives from HICPAC, the Society for Healthcare Epidemiology of America (SHEA), APIC, and the Infectious Diseases Society of America (IDSA).

Normal Bacterial Skin Flora

To understand the objectives of different approaches to hand cleansing, a knowledge of normal bacterial skin flora is essential. Normal human skin is colonized with bacteria; different areas of the body have varied total aerobic bacterial counts (e.g., 1 x 106 colony forming units (CFUs)/cm2 on the scalp, 5 x 105 CFUs/cm2 in the axilla, 4 x 104 CFUs/cm2 on the abdomen, and 1 x 104 CFUs/cm2 on the forearm) (13). Total bacterial counts on the hands of medical personnel have ranged from 3.9 x 104 to 4.6 x 106 (14--17). In 1938, bacteria recovered from the hands were divided into two categories: transient and resident (14). Transient flora, which colonize the superficial layers of the skin, are more amenable to removal by routine handwashing. They are often acquired by HCWs during direct contact with patients or contact with contaminated environmental surfaces within close proximity of the patient. Transient flora are the organisms most frequently associated with health-care--associated infections. Resident flora, which are attached to deeper layers of the skin, are more resistant to removal. In addition, resident flora (e.g., coagulase-negative staphylococci and diphtheroids) are less likely to be associated with such infections. The hands of HCWs may become persistently colonized with pathogenic flora (e.g., S. aureus), gram-negative bacilli, or yeast. Investigators have documented that, although the number of transient and resident flora varies considerably from person to person, it is often relatively constant for any specific person (14,18).

Physiology of Normal Skin

The primary function of the skin is to reduce water loss, provide protection against abrasive action and microorganisms, and act as a permeability barrier to the environment. The basic structure of skin includes, from outer- to inner-most layer, the superficial region (i.e., the stratum corneum or horny layer, which is 10- to 20-µm thick), the viable epidermis (50- to 100-µm thick), the dermis (1- to 2-mm thick), and the hypodermis (1- to 2-mm thick). The barrier to percutaneous absorption lies within the stratum corneum, the thinnest and smallest compartment of the skin. The stratum corneum contains the corneocytes (or horny cells), which are flat, polyhedral-shaped nonnucleated cells, remnants of the terminally differentiated keratinocytes located in the viable epidermis. Corneocytes are composed primarily of insoluble bundled keratins surrounded by a cell envelope stabilized by cross-linked proteins and covalently bound lipid. Interconnecting the corneocytes of the stratum corneum are polar structures (e.g., corneodesmosomes), which contribute to stratum corneum cohesion.

The intercellular region of the stratum corneum is composed of lipid primarily generated from the exocytosis of lamellar bodies during the terminal differentiation of the keratinocytes. The intercellular lipid is required for a competent skin barrier and forms the only continuous domain. Directly under the stratum corneum is a stratified epidermis, which is composed primarily of 10--20 layers of keratinizing epithelial cells that are responsible for the synthesis of the stratum corneum. This layer also contains melanocytes involved in skin pigmentation; Langerhans cells, which are important for antigen presentation and immune responses; and Merkel cells, whose precise role in sensory reception has yet to be fully delineated. As keratinocytes undergo terminal differentiation, they begin to flatten out and assume the dimensions characteristic of the corneocytes (i.e., their diameter changes from 10--12 µm to 20--30 µm, and their volume increases by 10- to 20-fold). The viable epidermis does not contain a vascular network, and the keratinocytes obtain their nutrients from below by passive diffusion through the interstitial fluid.

The skin is a dynamic structure. Barrier function does not simply arise from the dying, degeneration, and compaction of the underlying epidermis. Rather, the processes of cornification and desquamation are intimately linked; synthesis of the stratum corneum occurs at the same rate as loss. Substantial evidence now confirms that the formation of the skin barrier is under homeostatic control, which is illustrated by the epidermal response to barrier perturbation by skin stripping or solvent extraction. Circumstantial evidence indicates that the rate of keratinocyte proliferation directly influences the integrity of the skin barrier. A general increase in the rate of proliferation results in a decrease in the time available for 1) uptake of nutrients (e.g., essential fatty acids), 2) protein and lipid synthesis, and 3) processing of the precursor molecules required for skin-barrier function. Whether chronic but quantitatively smaller increases in rate of epidermal proliferation also lead to changes in skin-barrier function remains unclear. Thus, the extent to which the decreased barrier function caused by irritants is caused by an increased epidermal proliferation also is unknown.

The current understanding of the formation of the stratum corneum has come from studies of the epidermal responses to perturbation of the skin barrier. Experimental manipulations that disrupt the skin barrier include 1) extraction of skin lipids with apolar solvents, 2) physical stripping of the stratum corneum using adhesive tape, and 3) chemically induced irritation. All of these experimental manipulations lead to a decreased skin barrier as determined by transepidermal water loss (TEWL). The most studied experimental system is the treatment of mouse skin with acetone. This experiment results in a marked and immediate increase in TEWL, and therefore a decrease in skin-barrier function. Acetone treatment selectively removes glycerolipids and sterols from the skin, which indicates that these lipids are necessary, though perhaps not sufficient in themselves, for barrier function. Detergents act like acetone on the intercellular lipid domain. The return to normal barrier function is biphasic: 50%--60% of barrier recovery typically occurs within 6 hours, but complete normalization of barrier function requires 5--6 days.

Definition of Terms

Alcohol-based hand rub. An alcohol-containing preparation designed for application to the hands for reducing the number of viable microorganisms on the hands. In the United States, such preparations usually contain 60%--95% ethanol or isopropanol.

 Antimicrobial soap. Soap (i.e., detergent) containing an antiseptic agent.

Antiseptic agent. Antimicrobial substances that are applied to the skin to reduce the number of microbial flora. Examples include alcohols, chlorhexidine, chlorine, hexachlorophene, iodine, chloroxylenol (PCMX), quaternary ammonium compounds, and triclosan.

Antiseptic handwash. Washing hands with water and soap or other detergents containing an antiseptic agent.

Antiseptic hand rub. Applying an antiseptic hand-rub product to all surfaces of the hands to reduce the number of microorganisms present.

Cumulative effect. A progressive decrease in the numbers of microorganisms recovered after repeated applications of a test material.

Decontaminate hands. To Reduce bacterial counts on hands by performing antiseptic hand rub or antiseptic handwash.

Detergent. Detergents (i.e., surfactants) are compounds that possess a cleaning action. They are composed of both hydrophilic and lipophilic parts and can be divided into four groups: anionic, cationic, amphoteric, and nonionic detergents. Although products used for handwashing or antiseptic handwash in health-care settings represent various types of detergents, the term "soap" is used to refer to such detergents in this guideline.

Hand antisepsis. Refers to either antiseptic handwash or antiseptic hand rub.

Hand hygiene. A general term that applies to either handwashing, antiseptic handwash, antiseptic hand rub, or surgical hand antisepsis.

Handwashing. Washing hands with plain (i.e., non-antimicrobial) soap and water.

Persistent activity. Persistent activity is defined as the prolonged or extended antimicrobial activity that prevents or inhibits the proliferation or survival of microorganisms after application of the product. This activity may be demonstrated by sampling a site several minutes or hours after application and demonstrating bacterial antimicrobial effectiveness when compared with a baseline level. This property also has been referred to as "residual activity." Both substantive and nonsubstantive active ingredients can show a persistent effect if they substantially lower the number of bacteria during the wash period.

Plain soap. Plain soap refers to detergents that do not contain antimicrobial agents or contain low concentrations of antimicrobial agents that are effective solely as preservatives.

Substantivity. Substantivity is an attribute of certain active ingredients that adhere to the stratum corneum (i.e., remain on the skin after rinsing or drying) to provide an inhibitory effect on the growth of bacteria remaining on the skin.

Surgical hand antisepsis. Antiseptic handwash or antiseptic hand rub performed preoperatively by surgical personnel to eliminate transient and reduce resident hand flora. Antiseptic detergent preparations often have persistent antimicrobial activity.

Visibly soiled hands. Hands showing visible dirt or visibly contaminated with proteinaceous material, blood, or other body fluids (e.g., fecal material or urine).

Waterless antiseptic agent. An antiseptic agent that does not require use of exogenous water. After applying such an agent, the hands are rubbed together until the agent has dried.

Food and Drug Administration (FDA) product categories. The 1994 FDA Tentative Final Monograph for Health-Care Antiseptic Drug Products divided products into three categories and defined them as follows (19):

  • Patient preoperative skin preparation. A fast-acting, broad-spectrum, and persistent antiseptic-containing preparation that substantially reduces the number of microorganisms on intact skin.
  • Antiseptic handwash or HCW handwash. An antiseptic-containing preparation designed for frequent use; it reduces the number of microorganisms on intact skin to an initial baseline level after adequate washing, rinsing, and drying; it is broad-spectrum, fast-acting, and if possible, persistent.
  • Surgical hand scrub. An antiseptic-containing preparation that substantially reduces the number of microorganisms on intact skin; it is broad-spectrum, fast-acting, and persistent.

Evidence of Transmission of Pathogens on Hands

Transmission of health-care--associated pathogens from one patient to another via the hands of HCWs requires the following sequence of events:

  • Organisms present on the patient's skin, or that have been shed onto inanimate objects in close proximity to the patient, must be transferred to the hands of HCWs.
  • These organisms must then be capable of surviving for at least several minutes on the hands of personnel.
  • Next, handwashing or hand antisepsis by the worker must be inadequate or omitted entirely, or the agent used for hand hygiene must be inappropriate.
  • Finally, the contaminated hands of the caregiver must come in direct contact with another patient, or with an inanimate object that will come into direct contact with the patient.

Health-care--associated pathogens can be recovered not only from infected or draining wounds, but also from frequently colonized areas of normal, intact patient skin (20-- 31). The perineal or inguinal areas are usually most heavily colonized, but the axillae, trunk, and upper extremities (including the hands) also are frequently colonized (23,25,26,28,30--32). The number of organisms (e.g., S. aureus, Proteus mirabilis, Klebsiella spp., and Acinetobacter spp.) present on intact areas of the skin of certain patients can vary from 100 to 106/cm2 (25,29,31,33). Persons with diabetes, patients undergoing dialysis for chronic renal failure, and those with chronic dermatitis are likely to have areas of intact skin that are colonized with S. aureus (34--41). Because approximately 106 skin squames containing viable microorganisms are shed daily from normal skin (42), patient gowns, bed linen, bedside furniture, and other objects in the patient's immediate environment can easily become contaminated with patient flora (30,43--46). Such contamination is particularly likely to be caused by staphylococci or enterococci, which are resistant to dessication.

Data are limited regarding the types of patient-care activities that result in transmission of patient flora to the hands of personnel (26,45--51). In the past, attempts have been made to stratify patient-care activities into those most likely to cause hand contamination (52), but such stratification schemes were never validated by quantifying the level of bacterial contamination that occurred. Nurses can contaminate their hands with 100--1,000 CFUs of Klebsiella spp. during "clean" activities (e.g., lifting a patient; taking a patient's pulse, blood pressure, or oral temperature; or touching a patient's hand, shoulder, or groin) (48). Similarly, in another study, hands were cultured of nurses who touched the groins of patients heavily colonized with P. mirabilis (25); 10--600 CFUs/mL of this organism were recovered from glove juice samples from the nurses' hands. Recently, other researchers studied contamination of HCWs' hands during activities that involved direct patient-contact wound care, intravascular catheter care, respiratory-tract care, and the handling of patient secretions (51). Agar fingertip impression plates were used to culture bacteria; the number of bacteria recovered from fingertips ranged from 0 to 300 CFUs. Data from this study indicated that direct patient contact and respiratory-tract care were most likely to contaminate the fingers of caregivers. Gram-negative bacilli accounted for 15% of isolates and S. aureus for 11%. Duration of patient-care activity was strongly associated with the intensity of bacterial contamination of HCWs' hands.

HCWs can contaminate their hands with gram-negative bacilli, S. aureus, enterococci, or Clostridium difficile by performing "clean procedures" or touching intact areas of the skin of hospitalized patients (26,45,46,53). Furthermore, personnel caring for infants with respiratory syncytial virus (RSV) infections have acquired RSV by performing certain activities (e.g., feeding infants, changing diapers, and playing with infants) (49). Personnel who had contact only with surfaces contaminated with the infants' secretions also acquired RSV by contaminating their hands with RSV and inoculating their oral or conjunctival mucosa. Other studies also have documented that HCWs may contaminate their hands (or gloves) merely by touching inanimate objects in patient rooms (46,53--56). None of the studies concerning hand contamination of hospital personnel were designed to determine if the contamination resulted in transmission of pathogens to susceptible patients.

Other studies have documented contamination of HCWs' hands with potential health-care--associated pathogens, but did not relate their findings to the specific type of preceding patient contact (15,17,57--62). For example, before glove use was common among HCWs, 15% of nurses working in an isolation unit carried a median of 1 x 104 CFUs of S. aureus on their hands (61). Of nurses working in a general hospital, 29% had S. aureus on their hands (median count: 3,800 CFUs), whereas 78% of those working in a hospital for dermatology patients had the organism on their hands (median count: 14.3 x 106 CFUs). Similarly, 17%--30% of nurses carried gram-negative bacilli on their hands (median counts: 3,400--38,000 CFUs). One study found that S. aureus could be recovered from the hands of 21% of intensive-care--unit personnel and that 21% of physician and 5% of nurse carriers had >1,000 CFUs of the organism on their hands (59). Another study found lower levels of colonization on the hands of personnel working in a neurosurgery unit, with an average of 3 CFUs of S. aureus and 11 CFUs of gram-negative bacilli (16). Serial cultures revealed that 100% of HCWs carried gram-negative bacilli at least once, and 64% carried S. aureus at least once.

Models of Hand Transmission

Several investigators have studied transmission of infectious agents by using different experimental models. In one study, nurses were asked to touch the groins of patients heavily colonized with gram-negative bacilli for 15 seconds --- as though they were taking a femoral pulse (25). Nurses then cleaned their hands by washing with plain soap and water or by using an alcohol hand rinse. After cleaning their hands, they touched a piece of urinary catheter material with their fingers, and the catheter segment was cultured. The study revealed that touching intact areas of moist skin of the patient transferred enough organisms to the nurses' hands to result in subsequent transmission to catheter material, despite handwashing with plain soap and water.

The transmission of organisms from artificially contaminated "donor" fabrics to clean "recipient" fabrics via hand contact also has been studied. Results indicated that the number of organisms transmitted was greater if the donor fabric or the hands were wet upon contact (63). Overall, only 0.06% of the organisms obtained from the contaminated donor fabric were transferred to recipient fabric via hand contact. Staphylococcus saprophyticus, Pseudomonas aeruginosa, and Serratia spp. were also transferred in greater numbers than was Escherichia coli from contaminated fabric to clean fabric after hand contact (64). Organisms are transferred to various types of surfaces in much larger numbers (i.e., >104) from wet hands than from hands that are thoroughly dried (65).

Relation of Hand Hygiene and Acquisition of Health-Care--Associated Pathogens

Hand antisepsis reduces the incidence of health-care--associated infections (66,67). An intervention trial using historical controls demonstrated in 1847 that the mortality rate among mothers who delivered in the First Obstetrics Clinic at the General Hospital of Vienna was substantially lower when hospital staff cleaned their hands with an antiseptic agent than when they washed their hands with plain soap and water (3).

In the 1960s, a prospective, controlled trial sponsored by the National Institutes of Health and the Office of the Surgeon General demonstrated that infants cared for by nurses who did not wash their hands after handling an index infant colonized with S. aureus acquired the organism more often and more rapidly than did infants cared for by nurses who used hexachlorophene to clean their hands between infant contacts (68). This trial provided evidence that, when compared with no handwashing, washing hands with an antiseptic agent between patient contacts reduces transmission of health-care--associated pathogens.

Trials have studied the effects of handwashing with plain soap and water versus some form of hand antisepsis on health-care--associated infection rates (69,70). Health-care--associated infection rates were lower when antiseptic handwashing was performed by personnel (69). In another study, antiseptic handwashing was associated with lower health-care--associated infection rates in certain intensive-care units, but not in others (70).

Health-care--associated infection rates were lower after antiseptic handwashing using a chlorhexidine-containing detergent compared with handwashing with plain soap or use of an alcohol-based hand rinse (71). However, because only a minimal amount of the alcohol rinse was used during periods when the combination regimen also was in use and because adherence to policies was higher when chlorhexidine was available, determining which factor (i.e., the hand-hygiene regimen or differences in adherence) accounted for the lower infection rates was difficult. Investigators have determined also that health-care--associated acquisition of MRSA was reduced when the antimicrobial soap used for hygienic handwashing was changed (72,73).

Increased handwashing frequency among hospital staff has been associated with decreased transmission of Klebsiella spp. among patients (48); these studies, however, did not quantitate the level of handwashing among personnel. In a recent study, the acquisition of various health-care--associated pathogens was reduced when hand antisepsis was performed more frequently by hospital personnel (74); both this study and another (75) documented that the prevalence of health-care--associated infections decreased as adherence to recommended hand-hygiene measures improved.

Outbreak investigations have indicated an association between infections and understaffing or overcrowding; the association was consistently linked with poor adherence to hand hygiene. During an outbreak investigation of risk factors for central venous catheter-associated bloodstream infections (76), after adjustment for confounding factors, the patient-to-nurse ratio remained an independent risk factor for bloodstream infection, indicating that nursing staff reduction below a critical threshold may have contributed to this outbreak by jeopardizing adequate catheter care. The understaffing of nurses can facilitate the spread of MRSA in intensive-care settings (77) through relaxed attention to basic control measures (e.g., hand hygiene). In an outbreak of Enterobacter cloacae in a neonatal intensive-care unit (78), the daily number of hospitalized children was above the maximum capacity of the unit, resulting in an available space per child below current recommendations. In parallel, the number of staff members on duty was substantially less than the number necessitated by the workload, which also resulted in relaxed attention to basic infection-control measures. Adherence to hand-hygiene practices before device contact was only 25% during the workload peak, but increased to 70% after the end of the understaffing and overcrowding period. Surveillance documented that being hospitalized during this period was associated with a fourfold increased risk of acquiring a health-care--associated infection. This study not only demonstrates the association between workload and infections, but it also highlights the intermediate cause of antimicrobial spread: poor adherence to hand-hygiene policies.

Methods Used To Evaluate the Efficacy of Hand-Hygiene Products

Current Methods

Investigators use different methods to study the in vivoefficacy of handwashing, antiseptic handwash, and surgical hand antisepsis protocols. Differences among the various studies include 1) whether hands are purposely contaminated with bacteria before use of test agents, 2) the method used to contaminate fingers or hands, 3) the volume of hand-hygiene product applied to the hands, 4) the time the product is in contact with the skin, 5) the method used to recover bacteria from the skin after the test solution has been used, and 6) the method of expressing the efficacy of the product (i.e., either percent reduction in bacteria recovered from the skin or log reduction of bacteria released from the skin). Despite these differences, the majority of studies can be placed into one of two major categories: studies focusing on products to remove transient flora and studies involving products that are used to remove resident flora from the hands. The majority of studies of products for removing transient flora from the hands of HCWs involve artificial contamination of the volunteer's skin with a defined inoculum of a test organism before the volunteer uses a plain soap, an antimicrobial soap, or a waterless antiseptic agent. In contrast, products tested for the preoperative cleansing of surgeons' hands (which must comply with surgical hand-antisepsis protocols) are tested for their ability to remove resident flora from without artificially contaminating the volunteers' hands.

In the United States, antiseptic handwash products intended for use by HCWs are regulated by FDA's Division of Over-the-Counter Drug Products (OTC). Requirements for in vitro and in vivo testing of HCW handwash products and surgical hand scrubs are outlined in the FDA Tentative Final Monograph for Healthcare Antiseptic Drug Products (TFM) (19). Products intended for use as HCW handwashes are evaluated by using a standardized method (19). Tests are performed in accordance with use directions for the test material. Before baseline bacterial sampling and before each wash with the test material, 5 mL of a standardized suspension of Serratia marcescens are applied to the hands and then rubbed over the surfaces of the hands. A specified volume of the test material is dispensed into the hands and is spread over the hands and lower one third of the forearms. A small amount of tap water is added to the hands, and hands are completely lathered for a specified time, covering all surfaces of the hands and the lower third of the forearms. Volunteers then rinse hands and forearms under 40ºC tap water for 30 seconds. Ten washes with the test formulation are required. After the first, third, seventh, and tenth washes, rubber gloves or polyethylene bags used for sampling are placed on the right and left hands, and 75 mL of sampling solution is added to each glove; gloves are secured above the wrist. All surfaces of the hand are massaged for 1 minute, and samples are obtained aseptically for quantitative culture. No neutralizer of the antimicrobial is routinely added to the sampling solution, but if dilution of the antimicrobial in the sampling fluid does not result in demonstrable neutralization, a neutralizer specific for the test formulation is added to the sampling solution. For waterless formulations, a similar procedure is used. TFM criteria for efficacy are as follows: a 2-log10 reduction of the indicator organism on each hand within 5 minutes after the first use, and a 3-log10 reduction of the indicator organism on each hand within 5 minutes after the tenth use (19).

Products intended for use as surgical hand scrubs have been evaluated also by using a standardized method (19). Volunteers clean under fingernails with a nail stick and clip their fingernails. All jewelry is removed from hands and arms. Hands and two thirds of forearms are rinsed with tap water (38ºC--42ºC) for 30 seconds, and then they are washed with a non-antimicrobial soap for 30 seconds and are rinsed for 30 seconds under tap water. Baseline microbial hand counts can then be determined. Next, a surgical scrub is performed with the test formulation using directions provided by the manufacturer. If no instructions are provided with the formulation, two 5-minute scrubs of hands and forearms followed by rinsing are performed. Reduction from baseline microbial hand counts is determined in a series of 11 scrubs conducted during 5 days. Hands are sampled at 1 minute, 3 hours, and 6 hours after the first scrubs on day 1, day 2, and day 5. After washing, volunteers wear rubber gloves; 75 mL of sampling solution are then added to one glove, and all surfaces of the hands are massaged for 1 minute. Samples are then taken aseptically and cultured quantitatively. The other glove remains on the other hand for 6 hours and is sampled in the same manner. TFM requires that formulations reduce the number of bacteria 1 log10 on each hand within 1 minute of product application and that the bacterial cell count on each hand does not subsequently exceed baseline within 6 hours on day 1; the formulation must produce a 2-log10 reduction in microbial flora on each hand within 1 minute of product application by the end of the second day of enumeration and a 3-log10 reduction of microbial flora on each hand within 1 minute of product use by the end of the fifth day when compared with the established baseline (19).

The method most widely used in Europe to evaluate the efficacy of hand-hygiene agents is European Standard 1500--1997 (EN 1500---Chemical disinfectants and antiseptics. Hygienic hand-rub test method and requirements) (79). This method requires 12--15 test volunteers and an 18- to 24-hour growth of broth culture of E. coli K12. Hands are washed with a soft soap, dried, and then immersed halfway to the metacarpals in the broth culture for 5 seconds. Hands are removed from the broth culture, excess fluid is drained off, and hands are dried in the air for 3 minutes. Bacterial recovery for the initial value is obtained by kneading the fingertips of each hand separately for 60 seconds in 10 mL of tryptic soy broth (TSB) without neutralizers. The hands are removed from the broth and disinfected with 3 mL of the hand-rub agent for 30 seconds in a set design. The same operation is repeated with total disinfection time not exceeding 60 seconds. Both hands are rinsed in running water for 5 seconds and water is drained off. Fingertips of each hand are kneaded separately in 10 mL of TSB with added neutralizers. These broths are used to obtain the final value. Log10 dilutions of recovery medium are prepared and plated out. Within 3 hours, the same volunteers are tested with the reference disinfectant (60% 2-propanol [isopropanol]) and the test product. Colony counts are performed after 24 and 48 hours of incubation at 36ºC. The average colony count of both left and right hand is used for evaluation. The log-reduction factor is calculated and compared with the initial and final values. The reduction factor of the test product should be superior or the same as the reference alcohol-based rub for acceptance. If a difference exists, then the results are analyzed statistically using the Wilcoxon test. Products that have log reductions substantially less than that observed with the reference alcohol-based hand rub (i.e., approximately 4 log10 reduction) are classified as not meeting the standard.

Because of different standards for efficacy, criteria cited in FDA TFM and the European EN 1500 document for establishing alcohol-based hand rubs vary (1,19,79). Alcohol-based hand rubs that meet TFM criteria for efficacy may not necessarily meet the EN 1500 criteria for efficacy (80). In addition, scientific studies have not established the extent to which counts of bacteria or other microorganisms on the hands need to be reduced to minimize transmission of pathogens in health-care facilities (1,8); whether bacterial counts on the hands must be reduced by 1 log10 (90% reduction), 2 log10 (99%), 3 log10 (99.9%), or 4 log10 (99.99%) is unknown. Several other methods also have been used to measure the efficacy of antiseptic agents against various viral pathogens (81--83).

Shortcomings of Traditional Methodologies

Accepted methods of evaluating hand-hygiene products intended for use by HCWs require that test volunteers wash their hands with a plain or antimicrobial soap for 30 seconds or 1 minute, despite the observation in the majority of studies that the average duration of handwashing by hospital personnel is <15 seconds (52,84--89). A limited number of investigators have used 15-second handwashing or hygienic hand-wash protocols (90--94). Therefore, almost no data exist regarding the efficacy of plain or antimicrobial soaps under conditions in which they are actually used by HCWs. Similarly, certain accepted methods for evaluating waterless antiseptic agents for use as antiseptic hand rubs require that 3 mL of alcohol be rubbed into the hands for 30 seconds, followed by a repeat application for the same duration. This type of protocol also does not reflect actual usage patterns among HCWs. Furthermore, volunteers used in evaluations of products are usually surrogates for HCWs, and their hand flora may not reflect flora found on the hands of personnel working in health-care settings. Further studies should be conducted among practicing HCWs using standardized protocols to obtain more realistic views of microbial colonization and risk of bacterial transfer and cross-transmission (51).

Review of Preparations Used for Hand Hygiene

Plain (Non-Antimicrobial) Soap

Soaps are detergent-based products that contain esterified fatty acids and sodium or potassium hydroxide. They are available in various forms including bar soap, tissue, leaflet, and liquid preparations. Their cleaning activity can be attributed to their detergent properties, which result in removal of dirt, soil, and various organic substances from the hands. Plain soaps have minimal, if any, antimicrobial activity. However, handwashing with plain soap can remove loosely adherent transient flora. For example, handwashing with plain soap and water for 15 seconds reduces bacterial counts on the skin by 0.6--1.1 log10, whereas washing for 30 seconds reduces counts by 1.8--2.8 log10 (1). However, in several studies, handwashing with plain soap failed to remove pathogens from the hands of hospital personnel (25,45). Handwashing with plain soap can result in paradoxical increases in bacterial counts on the skin (92,95--97). Non-antimicrobial soaps may be associated with considerable skin irritation and dryness (92,96,98), although adding emollients to soap preparations may reduce their propensity to cause irritation. Occasionally, plain soaps have become contaminated, which may lead to colonization of hands of personnel with gram-negative bacilli (99).


The majority of alcohol-based hand antiseptics contain either isopropanol, ethanol, n-propanol, or a combination of two of these products. Although n-propanol has been used in alcohol-based hand rubs in parts of Europe for many years, it is not listed in TFM as an approved active agent for HCW handwashes or surgical hand-scrub preparations in the United States. The majority of studies of alcohols have evaluated individual alcohols in varying concentrations. Other studies have focused on combinations of two alcohols or alcohol solutions containing limited amounts of hexachlorophene, quaternary ammonium compounds, povidone-iodine, triclosan, or chlorhexidine gluconate (61,93,100--119).

The antimicrobial activity of alcohols can be attributed to their ability to denature proteins (120). Alcohol solutions containing 60%--95% alcohol are most effective, and higher concentrations are less potent (120--122) because proteins are not denatured easily in the absence of water (120). The alcohol content of solutions may be expressed as percent by weight (w/w), which is not affected by temperature or other variables, or as percent by volume (vol/vol), which can be affected by temperature, specific gravity, and reaction concentration (123). For example, 70% alcohol by weight is equivalent to 76.8% by volume if prepared at 15ºC, or 80.5% if prepared at 25ºC (123). Alcohol concentrations in antiseptic hand rubs are often expressed as percent by volume (19).

Alcohols have excellent in vitro germicidal activity against gram-positive and gram-negative vegetative bacteria, including multidrug-resistant pathogens (e.g., MRSA and VRE), Mycobacterium tuberculosis,and various fungi (120--122,124--129). Certain enveloped (lipophilic) viruses (e.g., herpes simplex virus, human immunodeficiency virus [HIV], influenza virus, respiratory syncytial virus, and vaccinia virus) are susceptible to alcohols when tested in vitro (120,130,131) (Table 1). Hepatitis B virus is an enveloped virus that is somewhat less susceptible but is killed by 60%--70% alcohol; hepatitis C virus also is likely killed by this percentage of alcohol (132). In a porcine tissue carrier model used to study antiseptic activity, 70% ethanol and 70% isopropanol were found to reduce titers of an enveloped bacteriophage more effectively than an antimicrobial soap containing 4% chlorhexidine gluconate (133). Despite its effectiveness against these organisms, alcohols have very poor activity against bacterial spores, protozoan oocysts, and certain nonenveloped (nonlipophilic) viruses.

Numerous studies have documented the in vivo antimicrobial activity of alcohols. Alcohols effectively reduce bacterial counts on the hands (14,121,125,134). Typically, log reductions of the release of test bacteria from artificially contaminated hands average 3.5 log10 after a 30-second application and 4.0--5.0 log10 after a 1-minute application (1). In 1994, the FDA TFM classified ethanol 60%--95% as a Category I agent (i.e., generally safe and effective for use in antiseptic handwash or HCW hand-wash products) (19). Although TFM placed isopropanol 70%--91.3% in category IIIE (i.e., insufficient data to classify as effective), 60% isopropanol has subsequently been adopted in Europe as the reference standard against which alcohol-based hand-rub products are compared (79). Alcohols are rapidly germicidal when applied to the skin, but they have no appreciable persistent (i.e., residual) activity. However, regrowth of bacteria on the skin occurs slowly after use of alcohol-based hand antiseptics, presumably because of the sublethal effect alcohols have on some of the skin bacteria (135,136). Addition of chlorhexidine, quaternary ammonium compounds, octenidine, or triclosan to alcohol-based solutions can result in persistent activity (1).

Alcohols, when used in concentrations present in alcohol-based hand rubs, also have in vivo activity against several nonenveloped viruses (Table 2). For example, 70% isopropanol and 70% ethanol are more effective than medicated soap or nonmedicated soap in reducing rotavirus titers on fingerpads (137,138). A more recent study using the same test methods evaluated a commercially available product containing 60% ethanol and found that the product reduced the infectivity titers of three nonenveloped viruses (i.e., rotavirus, adenovirus, and rhinovirus) by >3 logs (81). Other nonenveloped viruses such as hepatitis A and enteroviruses (e.g., poliovirus) may require 70%--80% alcohol to be reliably inactivated (82,139). However, both 70% ethanol and a 62% ethanol foam product with emollients reduced hepatitis A virus titers on whole hands or fingertips more than nonmedicated soap; both were equally as effective as antimicrobial soap containing 4% chlorhexidine gluconate in reducing reduced viral counts on hands (140). In the same study, both 70% ethanol and the 62% ethanol foam product demonstrated greater virucidal activity against poliovirus than either non-antimicrobial soap or a 4% chlorhexidine gluconate-containing soap (140). However, depending on the alcohol concentration, the amount of time that hands are exposed to the alcohol, and viral variant, alcohol may not be effective against hepatitis A and other nonlipophilic viruses. The inactivation of nonenveloped viruses is influenced by temperature, disinfectant-virus volume ratio, and protein load (141). Ethanol has greater activity against viruses than isopropanol. Further in vitro and in vivo studies of both alcohol-based formulations and antimicrobial soaps are warranted to establish the minimal level of virucidal activity that is required to interrupt direct contact transmission of viruses in health-care settings.

Alcohols are not appropriate for use when hands are visibly dirty or contaminated with proteinaceous materials. However, when relatively small amounts of proteinaceous material (e.g., blood) are present, ethanol and isopropanol may reduce viable bacterial counts on hands more than plain soap or antimicrobial soap (142).

Alcohol can prevent the transfer of health-care--associated pathogens (25,63,64). In one study, gram-negative bacilli were transferred from a colonized patient's skin to a piece of catheter material via the hands of nurses in only 17% of experiments after antiseptic hand rub with an alcohol-based hand rinse (25). In contrast, transfer of the organisms occurred in 92% of experiments after handwashing with plain soap and water. This experimental model indicates that when the hands of HCWs are heavily contaminated, an antiseptic hand rub using an alcohol-based rinse can prevent pathogen transmission more effectively than can handwashing with plain soap and water.

Alcohol-based products are more effective for standard handwashing or hand antisepsis by HCWs than soap or antimicrobial soaps (Table 3) (25,53,61,93,106--112,119,143--152). In all but two of the trials that compared alcohol-based solutions with antimicrobial soaps or detergents, alcohol reduced bacterial counts on hands more than washing hands with soaps or detergents containing hexachlorophene, povidone-iodine, 4% chlorhexidine, or triclosan. In studies examining antimicrobial-resistant organisms, alcohol-based products reduced the number of multidrug-resistant pathogens recovered from the hands of HCWs more effectively than did handwashing with soap and water (153--155).

Alcohols are effective for preoperative cleaning of the hands of surgical personnel (1,101,104,113--119,135,143,147,156--159) (Tables 4 and 5). In multiple studies, bacterial counts on the hands were determined immediately after using the product and again 1--3 hours later; the delayed testing was performed to determine if regrowth of bacteria on the hands is inhibited during operative procedures. Alcohol-based solutions were more effective than washing hands with plain soap in all studies, and they reduced bacterial counts on the hands more than antimicrobial soaps or detergents in the majority of experiments (101,104,113--119,135,143,147,157--159 ). In addition, the majority of alcohol-based preparations were more effective than povidone-iodine or chlorhexidine.

The efficacy of alcohol-based hand-hygiene products is affected by several factors, including the type of alcohol used, concentration of alcohol, contact time, volume of alcohol used, and whether the hands are wet when the alcohol is applied. Applying small volumes (i.e., 0.2--0.5 mL) of alcohol to the hands is not more effective than washing hands with plain soap and water (63,64). One study documented that 1 mL of alcohol was substantially less effective than 3 mL (91). The ideal volume of product to apply to the hands is not known and may vary for different formulations. However, if hands feel dry after rubbing hands together for 10--15 seconds, an insufficient volume of product likely was applied. Because alcohol-impregnated towelettes contain a limited amount of alcohol, their effectiveness is comparable to that of soap and water (63,160,161).

Alcohol-based hand rubs intended for use in hospitals are available as low viscosity rinses, gels, and foams. Limited data are available regarding the relative efficacy of various formulations. One field trial demonstrated that an ethanol gel was slightly more effective than a comparable ethanol solution at reducing bacterial counts on the hands of HCWs (162). However, a more recent study indicated that rinses reduced bacterial counts on the hands more than the gels tested (80). Further studies are warranted to determine the relative efficacy of alcohol-based rinses and gels in reducing transmission of health-care--associated pathogens.

Frequent use of alcohol-based formulations for hand antisepsis can cause drying of the skin unless emollients, humectants, or other skin-conditioning agents are added to the formulations. The drying effect of alcohol can be reduced or eliminated by adding 1%--3% glycerol or other skin-conditioning agents (90,93,100,101,106,135,143,163,164). Moreover, in several recent prospective trials, alcohol-based rinses or gels containing emollients caused substantially less skin irritation and dryness than the soaps or antimicrobial detergents tested (96,98,165,166). These studies, which were conducted in clinical settings, used various subjective and objective methods for assessing skin irritation and dryness. Further studies are warranted to establish whether products with different formulations yield similar results.

Even well-tolerated alcohol hand rubs containing emollients may cause a transient stinging sensation at the site of any broken skin (e.g., cuts and abrasions). Alcohol-based hand-rub preparations with strong fragrances may be poorly tolerated by HCWs with respiratory allergies. Allergic contact dermatitis or contact urticaria syndrome caused by hypersensitivity to alcohol or to various additives present in certain alcohol hand rubs occurs only rarely (167,168).

Alcohols are flammable. Flash points of alcohol-based hand rubs range from 21ºC to 24ºC, depending on the type and concentration of alcohol present (169). As a result, alcohol-based hand rubs should be stored away from high temperatures or flames in accordance with National Fire Protection Agency recommendations. In Europe, where alcohol-based hand rubs have been used extensively for years, the incidence of fires associated with such products has been low (169). One recent U.S. report described a flash fire that occurred as a result of an unusual series of events, which included an HCW applying an alcohol gel to her hands, immediately removing a polyester isolation gown, and then touching a metal door before the alcohol had evaporated (170). Removing the polyester gown created a substantial amount of static electricity that generated an audible static spark when the HCW touched the metal door, igniting the unevaporated alcohol on her hands (170). This incident emphasizes the need to rub hands together after application of alcohol-based products until all the alcohol has evaporated.

Because alcohols are volatile, containers should be designed to minimize evaporation. Contamination of alcohol-based solutions has seldom been reported. One report documented a cluster of pseudoinfections caused by contamination of ethyl alcohol by Bacillus cereus spores (171).


Chlorhexidine gluconate, a cationic bisbiguanide, was developed in England in the early 1950s and was introduced into the United States in the 1970s (8,172). Chlorhexidine base is only minimally soluble in water, but the digluconate form is water-soluble. The antimicrobial activity of chlorhexidine is likely attributable to attachment to, and subsequent disruption of, cytoplasmic membranes, resulting in precipitation of cellular contents (1,8). Chlorhexidine's immediate antimicrobial activity occurs more slowly than that of alcohols. Chlorhexidine has good activity against gram-positive bacteria, somewhat less activity against gram-negative bacteria and fungi, and only minimal activity against tubercle bacilli (1,8,172). Chlorhexidine is not sporicidal (1,172). It hasin vitroactivity against enveloped viruses (e.g., herpes simplex virus, HIV, cytomegalovirus, influenza, and RSV) but substantially less activity against nonenveloped viruses (e.g., rotavirus, adenovirus, and enteroviruses) (130,131,173). The antimicrobial activity of chlorhexidine is only minimally affected by the presence of organic material, including blood. Because chlorhexidine is a cationic molecule, its activity can be reduced by natural soaps, various inorganic anions, nonionic surfactants, and hand creams containing anionic emulsifying agents (8,172,174). Chlorhexidine gluconate has been incorporated into a number of hand-hygiene preparations. Aqueous or detergent formulations containing 0.5% or 0.75% chlorhexidine are more effective than plain soap, but they are less effective than antiseptic detergent preparations containing 4% chlorhexidine gluconate (135,175). Preparations with 2% chlorhexidine gluconate are slightly less effective than those containing 4% chlorhexidine (176).

Chlorhexidine has substantial residual activity (106,114--116,118,135,146,175). Addition of low concentrations (0.5%--1.0%) of chlorhexidine to alcohol-based preparations results in greater residual activity than alcohol alone (116,135). When used as recommended, chlorhexidine has a good safety record (172). Minimal, if any, absorption of the compound occurs through the skin. Care must be taken to avoid contact with the eyes when using preparations with >1% chlorhexidine, because the agent can cause conjunctivitis and severe corneal damage. Ototoxicity precludes its use in surgery involving the inner or middle ear. Direct contact with brain tissue and the meninges should be avoided. The frequency of skin irritation is concentration-dependent, with products containing 4% most likely to cause dermatitis when used frequently for antiseptic handwashing (177); allergic reactions to chlorhexidine gluconate are uncommon (118,172). Occasional outbreaks of nosocomial infections have been traced to contaminated solutions of chlorhexidine (178--181).


Chloroxylenol, also known as parachlorometaxylenol (PCMX), is a halogen-substituted phenolic compound that has been used as a preservative in cosmetics and other products and as an active agent in antimicrobial soaps. It was developed in Europe in the late 1920s and has been used in the United States since the 1950s (182).

The antimicrobial activity of PCMX likely is attributable to inactivation of bacterial enzymes and alteration of cell walls (1). It has good in vitro activity against gram-positive organisms and fair activity against gram-negative bacteria, mycobacteria, and certain viruses (1,7,182). PCMX is less active against P. aeruginosa, but addition of ethylene-diaminetetraacetic acid (EDTA) increases its activity against Pseudomonas spp. and other pathogens.

A limited number of articles focusing on the efficacy of PCMX-containing preparations intended for use by HCWs have been published in the last 25 years, and the results of studies have sometimes been contradictory. For example, in studies in which antiseptics were applied to abdominal skin, PCMX had the weakest immediate and residual activity of any of the agents studied (183). However, when 30-second handwashes were performed using 0.6% PCMX, 2% chlorhexidine gluconate, or 0.3% triclosan, the immediate effect of PCMX was similar to that of the other agents. When used 18 times per day for 5 consecutive days, PCMX had less cumulative activity than did chlorhexidine gluconate (184). When PCMX was used as a surgical scrub, one report indicated that 3% PCMX had immediate and residual activity comparable to 4% chlorhexidine gluconate (185), whereas two other studies demonstrated that the immediate and residual activity of PCMX was inferior to both chlorhexidine gluconate and povidone-iodine (176,186). The disparity between published studies may be associated with the various concentrations of PCMX included in the preparations evaluated and with other aspects of the formulations tested, including the presence or absence of EDTA (7,182). PCMX is not as rapidly active as chlorhexidine gluconate or iodophors, and its residual activity is less pronounced than that observed with chlorhexidine gluconate (7,182). In 1994, FDA TFM tentatively classified PCMX as a Category IIISE active agent (i.e., insufficient data are available to classify this agent as safe and effective) (19). Further evaluation of this agent by the FDA is ongoing.

The antimicrobial activity of PCMX is minimally affected by the presence of organic matter, but it is neutralized by nonionic surfactants. PCMX, which is absorbed through the skin (7,182), is usually well-tolerated, and allergic reactions associated with its use are uncommon. PCMX is available in concentrations of 0.3%--3.75%. In-use contamination of a PCMX-containing preparation has been reported (187).


Hexachlorophene is a bisphenol composed of two phenolic groups and three chlorine moieties. In the 1950s and early 1960s, emulsions containing 3% hexachlorophene were widely used for hygienic handwashing, as surgical scrubs, and for routine bathing of infants in hospital nurseries. The antimicrobial activity of hexachlorophene results from its ability to inactivate essential enzyme systems in microorganisms. Hexachlorophene is bacteriostatic, with good activity against S. aureus and relatively weak activity against gram-negative bacteria, fungi, and mycobacteria (7).

Studies of hexachlorophene as a hygienic handwash and surgical scrub demonstrated only modest efficacy after a single handwash (53,143,188). Hexachlorophene has residual activity for several hours after use and gradually reduces bacterial counts on hands after multiple uses (i.e., it has a cumulative effect) (1,101,188,189). With repeated use of 3% hexachlorophene preparations, the drug is absorbed through the skin. Infants bathed with hexachlorophene and personnel regularly using a 3% hexachlorophene preparation for handwashing have blood levels of 0.1--0.6 ppm hexachlorophene (190). In the early 1970s, certain infants bathed with hexachlorophene developed neurotoxicity (vacuolar degeneration) (191). As a result, in 1972, the FDA warned that hexachlorophene should no longer be used routinely for bathing infants. However, after routine use of hexachlorophene for bathing infants in nurseries was discontinued, investigators noted that the incidence of health-care--associated S. aureus infections in hospital nurseries increased substantially (192,193). In several instances, the frequency of infections decreased when hexachlorophene bathing of infants was reinstituted. However, current guidelines still recommend against the routine bathing of neonates with hexachlorophene because of its potential neurotoxic effects (194). The agent is classified by FDA TFM as not generally recognized as safe and effective for use as an antiseptic handwash (19). Hexachlorophene should not be used to bathe patients with burns or extensive areas of susceptible, sensitive skin. Soaps containing 3% hexachlorophene are available by prescription only (7).

Iodine and Iodophors

Iodine has been recognized as an effective antiseptic since the 1800s. However, because iodine often causes irritation and discoloring of skin, iodophors have largely replaced iodine as the active ingredient in antiseptics.

Iodine molecules rapidly penetrate the cell wall of microorganisms and inactivate cells by forming complexes with amino acids and unsaturated fatty acids, resulting in impaired protein synthesis and alteration of cell membranes (195). Iodophors are composed of elemental iodine, iodide or triiodide, and a polymer carrier (i.e., the complexing agent) of high molecular weight. The amount of molecular iodine present (so-called "free" iodine) determines the level of antimicrobial activity of iodophors. "Available" iodine refers to the total amount of iodine that can be titrated with sodium thiosulfate (196). Typical 10% povidone-iodine formulations contain 1% available iodine and yield free iodine concentrations of 1 ppm (196). Combining iodine with various polymers increases the solubility of iodine, promotes sustained release of iodine, and reduces skin irritation. The most common polymers incorporated into iodophors are polyvinyl pyrrolidone (i.e., povidone) and ethoxylated nonionic detergents (i.e., poloxamers) (195,196). The antimicrobial activity of iodophors also can be affected by pH, temperature, exposure time, concentration of total available iodine, and the amount and type of organic and inorganic compounds present (e.g., alcohols and detergents).

Iodine and iodophors have bactericidal activity against gram-positive, gram-negative, and certain spore-forming bacteria (e.g., clostridia and Bacillus spp.) and are active against mycobacteria, viruses, and fungi (8,195,197--200). However, in concentrations used in antiseptics, iodophors are not usually sporicidal (201). In vivo studies have demonstrated that iodophors reduce the number of viable organisms that are recovered from the hands of personnel (113,145,148,152,155). Povidone-iodine 5%--10% has been tentatively classified by FDA TFM as a Category I agent (i.e., a safe and effective agent for use as an antiseptic handwash and an HCW handwash) (19). The extent to which iodophors exhibit persistent antimicrobial activity after they have been washed off the skin is unclear. In one study, persistent activity was noted for 6 hours (176); however, several other studies demonstrated persistent activity for only 30--60 minutes after washing hands with an iodophor (61,117,202). In studies in which bacterial counts were obtained after gloves were worn for 1--4 hours after washing, iodophors have demonstrated poor persistent activity (1,104,115,189,203--208). The in vivo antimicrobial activity of iodophors is substantially reduced in the presence of organic substances (e.g., blood or sputum) (8).

The majority of iodophor preparations used for hand hygiene contain 7.5%--10% povidone-iodine. Formulations with lower concentrations also have good antimicrobial activity because dilution can increase free iodine concentrations (209). However, as the amount of free iodine increases, the degree of skin irritation also may increase (209). Iodophors cause less skin irritation and fewer allergic reactions than iodine, but more irritant contact dermatitis than other antiseptics commonly used for hand hygiene (92). Occasionally, iodophor antiseptics have become contaminated with gram-negative bacilli as a result of poor manufacturing processes and have caused outbreaks or pseudo-outbreaks of infection (196).

Quaternary Ammonium Compounds

Quaternary ammonium compounds are composed of a nitrogen atom linked directly to four alkyl groups, which may vary in their structure and complexity (210). Of this large group of compounds, alkyl benzalkonium chlorides are the most widely used as antiseptics. Other compounds that have been used as antiseptics include benzethonium chloride, cetrimide, and cetylpyridium chloride (1). The antimicrobial activity of these compounds was first studied in the early 1900s, and a quaternary ammonium compound for preoperative cleaning of surgeons' hands was used as early as 1935 (210). The antimicrobial activity of this group of compounds likely is attributable to adsorption to the cytoplasmic membrane, with subsequent leakage of low molecular weight cytoplasmic constituents (210).

Quaternary ammonium compounds are primarily bacteriostatic and fungistatic, although they are microbicidal against certain organisms at high concentrations (1); they are more active against gram-positive bacteria than against gram-negative bacilli. Quaternary ammonium compounds have relatively weak activity against mycobacteria and fungi and have greater activity against lipophilic viruses. Their antimicrobial activity is adversely affected by the presence of organic material, and they are not compatible with anionic detergents (1,210). In 1994, FDA TFM tentatively classified benzalkonium chloride and benzethonium chloride as Category IIISE active agents (i.e., insufficient data exists to classify them as safe and effective for use as an antiseptic handwash) (19). Further evaluation of these agents by FDA is in progress.

Quaternary ammonium compounds are usually well tolerated. However, because of weak activity against gram-negative bacteria, benzalkonium chloride is prone to contamination by these organisms. Several outbreaks of infection or pseudoinfection have been traced to quaternary ammonium compounds contaminated with gram-negative bacilli (211--213). For this reason, in the United States, these compounds have been seldom used for hand antisepsis during the last 15--20 years. However, newer handwashing products containing benzalkonium chloride or benzethonium chloride have recently been introduced for use by HCWs. A recent study of surgical intensive-care unit personnel found that cleaning hands with antimicrobial wipes containing a quaternary ammonium compound was about as effective as using plain soap and water for handwashing; both were less effective than decontaminating hands with an alcohol-based hand rub (214). One laboratory-based study reported that an alcohol-free hand-rub product containing a quaternary ammonium compound was efficacious in reducing microbial counts on the hands of volunteers (215). Further studies of such products are needed to determine if newer formulations are effective in health-care settings.


Triclosan (chemical name: 2,4,4' --trichloro-2'-hydroxy-diphenyl ether) is a nonionic, colorless substance that was developed in the 1960s. It has been incorporated into soaps for use by HCWs and the public and into other consumer products. Concentrations of 0.2%--2% have antimicrobial activity. Triclosan enters bacterial cells and affects the cytoplasmic membrane and synthesis of RNA, fatty acids, and proteins (216). Recent studies indicate this agent's antibacterial activity is attributable to binding to the active site of enoyl-acyl carrier protein reductase (217,218).

Triclosan has a broad range of antimicrobial activity, but it is often bacteriostatic (1). Minimum inhibitory concentrations (MICs) range from 0.1 to 10 ug/mL, whereas minimum bactericidal concentrations are 25--500 ug/mL. Triclosan's activity against gram-positive organisms (including MRSA) is greater than against gram-negative bacilli, particularly P. aeruginosa (1,216). The agent possesses reasonable activity against mycobacterial and Candida spp., but it has limited activity against filamentous fungi. Triclosan (0.1%) reduces bacterial counts on hands by 2.8 log10 after a 1-minute hygienic handwash (1). In several studies, log reductions have been lower after triclosan is used than when chlorhexidine, iodophors, or alcohol-based products are applied (1,61,149,184,219). In 1994, FDA TFM tentatively classified triclosan <1.0% as a Category IIISE active agent (i.e., insufficient data exist to classify this agent as safe and effective for use as an antiseptic handwash) (19). Further evaluation of this agent by the FDA is underway. Like chlorhexidine, triclosan has persistent activity on the skin. Its activity in hand-care products is affected by pH, the presence of surfactants, emollients, or humectants and by the ionic nature of the particular formulation (1,216). Triclosan's activity is not substantially affected by organic matter, but it can be inhibited by sequestration of the agent in micelle structures formed by surfactants present in certain formulations. The majority of formulations containing <2% triclosan are well-tolerated and seldom cause allergic reactions. Certain reports indicate that providing hospital personnel with a triclosan-containing preparation for hand antisepsis has led to decreased MRSA infections (72,73). Triclosan's lack of potent activity against gram-negative bacilli has resulted in occasional reports of contamination (220).

Other Agents

Approximately 150 years after puerperal-fever--related maternal mortality rates were demonstrated by Semmelweis to be reduced by use of a hypochlorite hand rinse, the efficacy of rubbing hands for 30 seconds with an aqueous hypochlorite solution was studied once again (221). The solution was demonstrated to be no more effective than distilled water. The regimen used by Semmelweis, which called for rubbing hands with a 4% [w/w] hypochlorite solution until the hands were slippery (approximately 5 minutes), has been revisited by other researchers (222). This more current study indicated that the regimen was 30 times more effective than a 1-minute rub using 60% isopropanol. However, because hypochlorite solutions are often irritating to the skin when used repeatedly and have a strong odor, they are seldom used for hand hygiene.

Certain other agents are being evaluated by FDA for use in health-care-related antiseptics (19). However, the efficacy of these agents has not been evaluated adequately for use in handwashing preparations intended for use by HCWs. Further evaluation of these agents is warranted. Products that use different concentrations of traditional antiseptics (e.g., low concentrations of iodophor) or contain novel compounds with antiseptic properties are likely to be introduced for use by HCWs. For example, preliminary studies have demonstrated that adding silver-containing polymers to an ethanol carrier (i.e., Surfacine®) results in a preparation that has persistent antimicrobial activity on animal and human skin (223). New compounds with good in vitro activity must be tested in vivo to determine their abilities to reduce transient and resident skin flora on the hands of HCWs.

Activity of Antiseptic Agents Against Spore-Forming Bacteria

The widespread prevalence of health-care--associated diarrhea caused by Clostridium difficile and the recent occurrence in the United States of human Bacillus anthracis infections associated with contaminated items sent through the postal system has raised concern regarding the activity of antiseptic agents against spore-forming bacteria. None of the agents (including alcohols, chlorhexidine, hexachlorophene, iodophors, PCMX, and triclosan) used in antiseptic handwash or antiseptic hand-rub preparations are reliably sporicidal against Clostridium spp. or Bacillus spp. (120,172,224,225). Washing hands with non-antimicrobial or antimicrobial soap and water may help to physically remove spores from the surface of contaminated hands. HCWs should be encouraged to wear gloves when caring for patients with C. difficile-associated diarrhea (226). After gloves are removed, hands should be washed with a non-antimicrobial or an antimicrobial soap and water or disinfected with an alcohol-based hand rub. During outbreaks of C. difficile-related infections, washing hands with a non-antimicrobial or antimicrobial soap and water after removing gloves is prudent. HCWs with suspected or documented exposure to B. anthracis-contaminated items also should be encouraged to wash their hands with a non-antimicrobial or antimicrobial soap and water.

Reduced Susceptibility of Bacteria to Antiseptics

Reduced susceptibility of bacteria to antiseptic agents can either be an intrinsic characteristic of a species or can be an acquired trait (227). Several reports have described strains of bacteria that appear to have acquired reduced susceptibility (when defined by MICs established in vitro) to certain antiseptics (e.g., chlorhexidine, quaternary ammonium compounds, and triclosan) (227--230). However, because the antiseptic concentrations that are actually used by HCWs are often substantially higher than the MICs of strains with reduced antiseptic susceptibility, the clinical relevance of the in vitro findings is questionable. For example, certain strains of MRSA have chlorhexidine and quaternary ammonium compound MICs that are several-fold higher than methicillin-susceptible strains, and certain strains of S. aureus have elevated MICs to triclosan (227,228). However, such strains were readily inhibited by the concentrations of these antiseptics that are actually used by practicing HCWs (227,228). The description of a triclosan-resistant bacterial enzyme has raised the question of whether resistance to this agent may develop more readily than to other antiseptic agents (218). In addition, exposing Pseudomonas strains containing the MexAB-OprM efflux system to triclosan may select for mutants that are resistant to multiple antibiotics, including fluoroquinolones (230). Further studies are needed to determine whether reduced susceptibility to antiseptic agents is of epidemiologic significance and whether resistance to antiseptics has any influence on the prevalence of antibiotic-resistant strains (227).

Surgical Hand Antisepsis

Since the late 1800s, when Lister promoted the application of carbolic acid to the hands of surgeons before procedures, preoperative cleansing of hands and forearms with an antiseptic agent has been an accepted practice (231). Although no randomized, controlled trials have been conducted to indicate that surgical-site infection rates are substantially lower when preoperative scrubbing is performed with an antiseptic agent rather than a non-antimicrobial soap, certain other factors provide a strong rationale for this practice. Bacteria on the hands of surgeons can cause wound infections if introduced into the operative field during surgery (232); rapid multiplication of bacteria occurs under surgical gloves if hands are washed with a non-antimicrobial soap. However, bacterial growth is slowed after preoperative scrubbing with an antiseptic agent (14,233). Reducing resident skin flora on the hands of the surgical team for the duration of a procedure reduces the risk of bacteria being released into the surgical field if gloves become punctured or torn during surgery (1,156,169). Finally, at least one outbreak of surgical-site infections occurred when surgeons who normally used an antiseptic surgical scrub preparation began using a non-antimicrobial product (234).

Antiseptic preparations intended for use as surgical hand scrubs are evaluated for their ability to reduce the number of bacteria released from hands at different times, including 1) immediately after scrubbing, 2) after wearing surgical gloves for 6 hours (i.e., persistent activity), and 3) after multiple applications over 5 days (i.e., cumulative activity). Immediate and persistent activity are considered the most important in determining the efficacy of the product. U.S. guidelines recommend that agents used for surgical hand scrubs should substantially reduce microorganisms on intact skin, contain a nonirritating antimicrobial preparation, have broad-spectrum activity, and be fast-acting and persistent (19,235).

Studies have demonstrated that formulations containing 60%--95% alcohol alone or 50%--95% when combined with limited amounts of a quaternary ammonium compound, hexachlorophene, or chlorhexidine gluconate, lower bacterial counts on the skin immediately postscrub more effectively than do other agents (Table 4). The next most active agents (in order of decreasing activity) are chlorhexidine gluconate, iodophors, triclosan, and plain soap (104,119,186,188, 203,204,206,208,236). Because studies of PCMX as a surgical scrub have yielded contradictory results, further studies are needed to establish how the efficacy of this compound compares with the other agents (176,185,186).

Although alcohols are not considered to have persistent antimicrobial activity, bacteria appear to reproduce slowly on the hands after a surgical scrub with alcohol, and bacterial counts on hands after wearing gloves for 1--3 hours seldom exceed baseline (i.e., prescrub) values (1). However, a recent study demonstrated that a formulation containing 61% ethanol alone did not achieve adequate persistent activity at 6 hours postscrub (237). Alcohol-based preparations containing 0.5% or 1% chlorhexidine gluconate have persistent activity that, in certain studies, has equaled or exceeded that of chlorhexidine gluconate-containing detergents (1,118,135,237).*

Persistent antimicrobial activity of detergent-based surgical scrub formulations is greatest for those containing 2% or 4% chlorhexidine gluconate, followed by hexachlorophene, triclosan, and iodophors (1,102,113--115,159,189,203, 204,206--208,236). Because hexachlorophene is absorbed into the blood after repeated use, it is seldom used as a surgical scrub.

Surgical staff have been traditionally required to scrub their hands for 10 minutes preoperatively, which frequently leads to skin damage. Several studies have demonstrated that scrubbing for 5 minutes reduces bacterial counts as effectively as a 10-minute scrub (117,238,239). In other studies, scrubbing for 2 or 3 minutes reduced bacterial counts to acceptable levels (156,205,207,240,241).

Studies have indicated that a two-stage surgical scrub using an antiseptic detergent, followed by application of an alcohol-containing preparation, is effective. For example, an initial 1- or 2-minute scrub with 4% chlorhexidine gluconate or povidone-iodine followed by application of an alcohol-based product has been as effective as a 5-minute scrub with an antiseptic detergent (114,242).

Surgical hand-antisepsis protocols have required personnel to scrub with a brush. But this practice can damage the skin of personnel and result in increased shedding of bacteria from the hands (95,243). Scrubbing with a disposable sponge or combination sponge-brush has reduced bacterial counts on the hands as effectively as scrubbing with a brush (244--246). However, several studies indicate that neither a brush nor a sponge is necessary to reduce bacterial counts on the hands of surgical personnel to acceptable levels, especially when alcohol-based products are used (102,117,159,165,233,237, 247,248). Several of these studies performed cultures immediately or at 45--60 minutes postscrub (102,117, 233,247,248), whereas in other studies, cultures were obtained 3 and 6 hours postscrub (159,237). For example, a recent laboratory-based study using volunteers demonstrated that brushless application of a preparation containing 1% chlorhexidine gluconate plus 61% ethanol yielded lower bacterial counts on the hands of participants than using a sponge/brush to apply a 4% chlorhexidine-containing detergent preparation (237).

Relative Efficacy of Plain Soap, Antiseptic Soap/Detergent, and Alcohols

Comparing studies related to the in vivo efficacy of plain soap, antimicrobial soaps, and alcohol-based hand rubs is problematic, because certain studies express efficacy as the percentage reduction in bacterial counts achieved, whereas others give log10 reductions in counts achieved. However, summarizing the relative efficacy of agents tested in each study can provide an overview of the in vivo activity of various formulations intended for handwashing, hygienic handwash, antiseptic hand rub, or surgical hand antisepsis (Tables 2--4).

Irritant Contact Dermatitis Resulting from Hand-Hygiene Measures

Frequency and Pathophysiology of Irritant Contact Dermatitis

In certain surveys, approximately 25% of nurses report symptoms or signs of dermatitis involving their hands, and as many as 85% give a history of having skin problems (249). Frequent and repeated use of hand-hygiene products, particularly soaps and other detergents, is a primary cause of chronic irritant contact dermatitis among HCWs (250). The potential of detergents to cause skin irritation can vary considerably and can be ameliorated by the addition of emollients and humectants. Irritation associated with antimicrobial soaps may be caused by the antimicrobial agent or by other ingredients of the formulation. Affected persons often complain of a feeling of dryness or burning; skin that feels "rough;" and erythema, scaling, or fissures. Detergents damage the skin by causing denaturation of stratum corneum proteins, changes in intercellular lipids (either depletion or reorganization of lipid moieties), decreased corneocyte cohesion, and decreased stratum corneum water-binding capacity (250,251). Damage to the skin also changes skin flora, resulting in more frequent colonization by staphylococci and gram-negative bacilli (17,90). Although alcohols are among the safest antiseptics available, they can cause dryness and irritation of the skin (1,252). Ethanol is usually less irritating than n-propanol or isopropanol (252).

Irritant contact dermatitis is more commonly reported with iodophors (92). Other antiseptic agents that can cause irritant contact dermatitis (in order of decreasing frequency) include chlorhexidine, PCMX, triclosan, and alcohol-based products. Skin that is damaged by repeated exposure to detergents may be more susceptible to irritation by alcohol-based preparations (253). The irritancy potential of commercially prepared hand-hygiene products, which is often determined by measuring transepidermal water loss, may be available from the manufacturer. Other factors that can contribute to dermatitis associated with frequent handwashing include using hot water for handwashing, low relative humidity (most common in winter months), failure to use supplementary hand lotion or cream, and the quality of paper towels (254,255). Shear forces associated with wearing or removing gloves and allergy to latex proteins may also contribute to dermatitis of the hands of HCWs.

Allergic Contact Dermatitis Associated with Hand-Hygiene Products

Allergic reactions to products applied to the skin (i.e., contact allergies) may present as delayed type reactions (i.e., allergic contact dermatitis) or less commonly as immediate reactions (i.e., contact urticaria). The most common causes of contact allergies are fragrances and preservatives; emulsifiers are less common causes (256--259). Liquid soaps, hand lotions or creams, and "udder ointments" may contain ingredients that cause contact allergies among HCWs (257,258).

Allergic reactions to antiseptic agents, including quaternary ammonium compounds, iodine or iodophors, chlorhexidine, triclosan, PCMX, and alcohols have been reported (118,167,172,256,260--265). Allergic contact dermatitis associated with alcohol-based hand rubs is uncommon. Surveillance at a large hospital in Switzerland, where a commercial alcohol hand rub has been used for >10 years, failed to identify a single case of documented allergy to the product (169). In late 2001, a Freedom of Information Request for data in the FDA's Adverse Event Reporting System regarding adverse reactions to popular alcohol hand rubs in the United States yielded only one reported case of an erythematous rash reaction attributed to such a product (John M. Boyce, M.D., Hospital of St. Raphael, New Haven, Connecticut, personal communication, 2001). However, with increasing use of such products by HCWs, true allergic reactions to such products likely will be encountered.

Allergic reactions to alcohol-based products may represent true allergy to alcohol, allergy to an impurity or aldehyde metabolite, or allergy to another constituent of the product (167). Allergic contact dermatitis or immediate contact urticarial reactions may be caused by ethanol or isopropanol (167). Allergic reactions can be caused by compounds that may be present as inactive ingredients in alcohol-based hand rubs, including fragrances, benzyl alcohol, stearyl or isostearyl alcohol, phenoxyethanol, myristyl alcohol, propylene glycol, parabens, and benzalkonium chloride (167,256,266--270).

Proposed Methods for Reducing Adverse Effects of Agents

Potential strategies for minimizing hand-hygiene--related irritant contact dermatitis among HCWs include reducing the frequency of exposure to irritating agents (particularly anionic detergents), replacing products with high irritation potential with preparations that cause less damage to the skin, educating personnel regarding the risks of irritant contact dermatitis, and providing caregivers with moisturizing skin-care products or barrier creams (96,98,251,271--273). Reducing the frequency of exposure of HCWs to hand-hygiene products would prove difficult and is not desirable because of the low levels of adherence to hand-hygiene policies in the majority of institutions. Although hospitals have provided personnel with non-antimicrobial soaps in hopes of minimizing dermatitis, frequent use of such products may cause greater skin damage, dryness, and irritation than antiseptic preparations (92,96,98). One strategy for reducing the exposure of personnel to irritating soaps and detergents is to promote the use of alcohol-based hand rubs containing various emollients. Several recent prospective, randomized trials have demonstrated that alcohol-based hand rubs containing emollients were better tolerated by HCWs than washing hands with non-antimicrobial soaps or antimicrobial soaps (96,98,166). Routinely washing hands with soap and water immediately after using an alcohol hand rub may lead to dermatitis. Therefore, personnel should be reminded that it is neither necessary nor recommended to routinely wash hands after each application of an alcohol hand rub.

Hand lotions and creams often contain humectants and various fats and oils that can increase skin hydration and replace altered or depleted skin lipids that contribute to the barrier function of normal skin (251,271). Several controlled trials have demonstrated that regular use (e.g., twice a day) of such products can help prevent and treat irritant contact dermatitis caused by hand-hygiene products (272,273). In one study, frequent and scheduled use of an oil-containing lotion improved skin condition, and thus led to a 50% increase in handwashing frequency among HCWs (273). Reports from these studies emphasize the need to educate personnel regarding the value of regular, frequent use of hand-care products.

Recently, barrier creams have been marketed for the prevention of hand-hygiene--related irritant contact dermatitis. Such products are absorbed to the superficial layers of the epidermis and are designed to form a protective layer that is not removed by standard handwashing. Two recent randomized, controlled trials that evaluated the skin condition of caregivers demonstrated that barrier creams did not yield better results than did the control lotion or vehicle used (272,273). As a result, whether barrier creams are effective in preventing irritant contact dermatitis among HCWs remains unknown.

In addition to evaluating the efficacy and acceptability of hand-care products, product-selection committees should inquire about the potential deleterious effects that oil-containing products may have on the integrity of rubber gloves and on the efficacy of antiseptic agents used in the facility (8,236).

Factors To Consider When Selecting Hand-Hygiene Products

When evaluating hand-hygiene products for potential use in health-care facilities, administrators or product-selection committees must consider factors that can affect the overall efficacy of such products, including the relative efficacy of antiseptic agents against various pathogens (Appendix) and acceptance of hand-hygiene products by personnel (274,275). Soap products that are not well-accepted by HCWs can be a deterrent to frequent handwashing (276). Characteristics of a product (either soap or alcohol-based hand rub) that can affect acceptance by personnel include its smell, consistency (i.e., "feel"), and color (92,277,278). For soaps, ease of lathering also may affect user preference.

Because HCWs may wash their hands from a limited number of times per shift to as many as 30 times per shift, the tendency of products to cause skin irritation and dryness is a substantial factor that influences acceptance, and ultimate usage (61,98,274,275,277,279). For example, concern regarding the drying effects of alcohol was a primary cause of poor acceptance of alcohol-based hand-hygiene products in hospitals in the United States (5,143). However, several studies have demonstrated that alcohol-based hand rubs containing emollients are acceptable to HCWs (90,93,98,100,101,106, 143,163,164,166). With alcohol-based products, the time required for drying may also affect user acceptance.

Studies indicate that the frequency of handwashing or antiseptic handwashing by personnel is affected by the accessibility of hand-hygiene facilities (280--283