Major article

Application of a theoretical framework for behavior change to hospital workers’ real-time explanations for noncompliance with hand hygiene guidelines

  • a UCL Research Department of Infection and Population Health, University College London, London, United Kingdom
  • b NIHR King’s Patient Safety and Service Quality Research Centre, King’s College London, London, United Kingdom
  • c Scottish Collaboration for Public Health Research and Policy, Medical Research Council, University of Edinburgh, Edinburgh, Scotland, United Kingdom
  • d UCL Medical School, University College London, London, United Kingdom
  • e Department of Clinical, Educational and Health Psychology, University College London, London, United Kingdom


Insufficient use of behavioral theory to understand health care workers’ (HCWs) hand hygiene compliance may result in suboptimal design of hand hygiene interventions and limit effectiveness. Previous studies examined HCWs’ intended, rather than directly observed, compliance and/or focused on just 1 behavioral model. This study examined HCWs’ explanations of noncompliance in “real time” (immediately after observation), using a behavioral theory framework, to inform future intervention design.


HCWs were directly observed and asked to explain episodes of noncompliance in “real-time.” Explanations were recorded, coded into 12 behavioral domains, using the Theory Domains Framework, and subdivided into themes.


Over two-thirds of 207 recorded explanations were explained by 2 domains. These were “Memory/Attention/Decision Making” (87, 44%), subdivided into 3 themes (memory, loss of concentration, and distraction by interruptions), and “Knowledge” (55, 26%), with 2 themes relating to specific hand hygiene indications. No other domain accounted for more than 18 (9%) explanations.


An explanation of HCW’s “real-time” explanations for noncompliance identified “Memory/Attention/Decision Making” and “Knowledge” as the 2 behavioral domains commonly linked to noncompliance. This suggests that hand hygiene interventions should target both automatic associative learning processes and conscious decision making, in addition to ensuring good knowledge. A theoretical framework to investigate HCW’s “real-time” explanations of noncompliance provides a coherent way to design hand hygiene interventions.

Key Words

  • Behavioral theory;
  • Noncompliance

Sustained improvements in hand hygiene compliance are central to the World Health Organization’s (WHO) First Global Patient Safety challenge to reduce the international burden of health care-associated infection.1 Systematic reviews of interventions to improve health care workers’ (HCWs) hand hygiene compliance show that improvement is difficult to achieve and sustain.2 and 3 This may be, in part, due to the failure of studies to use behavioral theory in their design.4 Understanding the barriers to and drivers of hand hygiene from a theory-based perspective would facilitate the design of an optimally effective intervention.5

A recent systematic review of HCWs’ compliance with hand hygiene guidelines in hospital6 suggested that insufficient use of behavioral theory has limited our understanding of hand hygiene behavior. This may have arisen for 2 reasons. First, previous studies have generally linked predictors of hand hygiene with HCWs’ intended or self-reported behavior rather than their actual directly observed behavior. This is despite the fact that there are well-recognized discrepancies between the two.7 and 8 Second, studies have often focused on one model of behavior, such as the Theory of Planned Behavior,9 although one theory or model is unlikely to encapsulate the wide range of influences on hand hygiene behavior cited in the literature.16

The Theoretical Domains Framework (TDF)10 and 11 is a well-validated, comprehensive, consensus-based, theoretical framework for understanding implementation of health care guidelines. It uses a core set of 12 theoretical domains based on psychologic theories and constructs relevant to implementation of evidence-based practice, accompanied by exemplar interview questions to elicit which domains are relevant to any given behavior. To improve design of future hand hygiene interventions, we used the TDF to analyze HCWs’ explanations of their noncompliant hand hygiene behavior recorded in “real time” (ie, shortly after direct observation).



A cross-sectional study was used, nested within a cluster randomized controlled trial of a hand hygiene intervention, the Feedback Intervention Trial (FIT) ( ISRCTN65246961). Ethical approval for the study was received from the Scotland B Multicentre Research Ethics Committee 05/MREC10/2.


Eleven intensive therapy units and 22 acute care units of the elderly/general medical wards implemented the FIT in 13 English and Welsh hospitals. The FIT intervention was informed by behavioral theory, specifically goal setting,12 control,13 and operant learning14 theories and used personalized goal setting and action planning to augment the effect of feedback. It involved a repeating 4-week audit cycle and was carried out by a designated member of staff, trained as a “ward coordinator.” Notices were posted on each ward explaining that hand hygiene observations were being carried out and that any member of staff might be observed. In weeks 1 and 2, the ward coordinator covertly observed an individual HCW for 20 minutes. Immediate feedback was given to the individual directly after the period of observation. When the HCWs were noncompliant with hand hygiene, they were helped to formulate an individual goal and action plan to improve hand hygiene. Ward coordinators filled out a form recording observations, feedback, goals, and action plans and returned these to the FIT trial manager (C.F.). Details of the trial, the feedback forms, and the training given to ward coordinators in hand hygiene observation, feedback and helping HCWs set goals and action plans are available elsewhere (, 16 and 17 Participants included HCWs on wards taking part in the FIT between September 2007 and August 2009.

Data collection

The feedback process included asking HCWs to explain their episodes of poor compliance and required the ward coordinators to document these “real-time” self-reported explanations of noncompliance on the feedback forms. This provided the current study, which was conducted after the conclusion of the trial, with a novel means of identifying self-reported barriers to hand hygiene that were linked in “real time” to an observed behavior.


In total, 570 feedback forms were returned, of which 209 indicated that hand hygiene had not been carried out. One hundred eighty-five (89%) forms recorded 1 or more reasons given for not cleaning hands, of which 119 of 185 (64%) were from observations carried out on nurses, 19 of 185 (10%) from doctors, 35 of 185 (19%) from other HCWs (ie, professions allied to medicine and ancillary staff), and 12 (7%) from unrecorded staff groups. The forms were completed by a total of 52 ward coordinators (usually a junior sister/deputy ward manager) or infection control link nurse.


Analysis was undertaken after the conclusion of the FIT trial. Formal standard operating procedures (available from the author) for the classification of responses into the 12 standardized behavioral domains (the Theory Domains Framework)5 were developed following discussion among the research team (C.F., S.B., J.S., J.M.). Explanations for noncompliance were then coded independently by 2 members of the FIT team (C.F. and S.B.). Disagreements were discussed, and any that were not resolved were discussed with a third researcher (J.S.) until consensus was reached. The percentage agreement between the 2 independent coders for all 185 forms was 85% (158/185). Of the 27 disagreements, 16 were agreed on discussion, 11 were resolved in consultation (with J.S.), and 3 were not resolved and coded as “Other.”

Following this, 2 members of the FIT team (C.F. and S.B.) independently examined the data to identify themes that could further divide HCWs explanations within individual domains. Both researchers discussed each theme until consensus was reached as to which themes accurately reflected the data. A coding framework was developed (available from the author), and each explanation was coded independently. Disagreements in coding were discussed until agreement was reached.


In total, 185 forms recording noncompliance with hand hygiene guidelines were completed. Of these, 22 recorded more than 1 explanation for noncompliance and were coded to more than 1 theoretical domain, leading to a total of 207 domain codings. Table 1 summarizes the numbers of explanations coded to each domain and theme, along with operational definitions and examples for each. All 207 explanations, broken down into their relevant domains and themes, are available from the author.