Mr Gordon Rushworth, Advanced Pharmacist Clinical Research, Highland Clinical Research Facility, Centre for Health Science, Old Perth Road, Inverness IV2 3JH, UK. E-mail: firstname.lastname@example.org
Objective To explore the association between medication adherence and qualitatively characterised patient-specific themes relating to medication adherence in patients following percutaneous coronary intervention (PCI).
Methods Data-collection questionnaires and qualitative topic guides were piloted in two patients. A validated questionnaire generated an adherence score for a convenience sample of 20 patients within 7 days of PCI. Semi-structured qualitative interviews were subsequently carried out with all patients to explore patient-specific themes relating to measured medication adherence.
Key findings Fourteen out of 20 patients (70%) had scores indicative of good adherence. Key factors associated with good adherence included having a good relationship with the doctor, having an understanding of the condition, knowledge of the indications and consequences of non-adherence, perceived health benefits and medications eliciting tangible symptom control. There were misconceptions of concern regarding adverse drug reactions and the importance of aspirin, both of which had a negative effect on adherence. The role of the community pharmacist was sometimes, although not always, misunderstood.
Conclusion This study suggests there is an association between patients' beliefs, knowledge, understanding and misconceptions about medication and their adherence in a post-PCI cohort. To optimise medication adherence it is vital for prescribers to remain patient-focused and cognisant of patient-specific themes relating to medication adherence.
The concept of patient adherence to medication is unique from compliance. Adherence avoids some of the negative connotations that the term ‘compliance’ has been tainted with over the last two millennia of traditional physician-to-patient consultations. Compliance is a simplistic term which relates to the degree to which the patient follows the direct instructions of the prescriber. Moreover, with the idea of adherence comes an additional concept related to understanding why patients are adherent, or otherwise. In turn, this enables differentiation between patients who have purposefully chosen not to take a medication (intentional non-adherence) and those that have not been able to take their medication due to practical reasons (unintentional non-adherence).[1–3] The key subtle difference between the two terms stems from the ability to understand why patients are not taking their prescribed medication. The benefits of this stratification are revealed when considering health-seeking behaviour. Recent guidance from the UK National Institute for Health and Clinical Excellence (NICE) has reiterated the importance of determining the rationale for a patient's decision to take, or not take, medication. This reasoning can then be explored to find a mutual solution to potential adherence problems.
In patients prescribed statins, non-adherence was influenced by patients' own beliefs about their medication and the perceived benefit derived from them. Beliefs about medication have been identified as being a predictor of adherence. A number of studies have defined the benefit(s) patients perceive that they will gain from their medication.[5,7–10] Therefore, in order to improve medication adherence it is essential to understand more about patients' beliefs regarding their medication. There is evidence that adherence may be enhanced by improving patient education and counselling. In taking this approach, healthcare professionals should be cognisant of the level of understanding patients may be able to achieve. Views regarding the benefits of medication should be discussed during the consultation, and at the point of prescribing between the prescriber and patient. Patients will be able to appreciate the benefits of their medication if they have better understanding, especially when they are required to take them for long periods of time.[9,13] Notably, misconceptions surrounding disease states are associated with poorer physical health; in turn, a poor understanding of the disease increases the likelihood that the patient will not understand the benefits of taking their medication.
Following percutaneous coronary intervention (PCI) patients fall under the auspices of being treated for a long-term condition – coronary heart disease – and therefore require medication. PCI can be done either electively or after an acute event. According to World Health Organization data, the average adherence rate for patients on medication for long-term conditions is 50%.
One retrospective study in patients prescribed antiplatelet therapy (clopidogrel) following an acute coronary syndrome (ACS) who had been treated with either a bare metal stent (BMS) or a drug-eluting stent (DES) reported higher mortality in patients who discontinued clopidogrel early; up to 18 months post-ACS. Furthermore, the increase in adverse events appears highest in the first 90 days after stopping the thienopyridine antiplatelet clopidogrel in both medically and PCI-treated ACS patients (incidence rate ratios 1.98 and 1.82 respectively). This study did not explore the reasons why patients stopped taking thienopyridine drug therapy. Even assuming that adherence to dual antiplatelet post-PCI medication is good, stent thrombosis occurs in 0.5–2% of elective and up to 6% of ACS patients who are given a stent. Thus the risk of a cardiovascular event due to stent thrombosis increases with increasing non-adherence.
In a further study investigating the prevalence and predictors of thienopyridine antiplatelet discontinuation post-myocardial infarction (MI) in patients treated with BMS, almost one in seven patients discontinued thienopyridine by day 30. This was associated with a significantly higher increase in mortality over the next 11 months (7.5 compared with 0.7%, P<0.0001). Those who discontinued were less educated, not married, had previous co-morbidities and were generally older. What the study did not illustrate, beyond interpretation of demographic data, were the reasons why individual patients had stopped their medication. However, it does allow for hypotheses to be drawn from the results, which can be explored further using qualitative techniques.
The effect of medication cost in relation to adherence has been studied by Ko et al. in 10 000 patients, all of whom were above the age of 65 and had received either BMS or DES as PCI in Canada. Thienopyridine antiplatelet therapy was given to patients at low cost. This study found that non-adherence was highest in the patients who had to pay the most for their prescription. The group who received free medication were almost 70% more likely to order prescriptions, thus implying a prohibitive effect of healthcare charges and supporting the argument that patients who have to pay for medication are less likely to access it. Non-adherence increased the risk of mortality. The investigators also found that patient adherence decreased with increasing time after the index event, suggesting that a degree of ambivalence manifests with time.
The effect of adherence to statin therapy has also been investigated post-PCI. The relative risk reduction for those on statin post-PCI was reported as 22% in the original trial. After analysis and adjusting for non-compliance, the relative risk reduction for major cardiac events was 32%, with the additional 10% relative risk reduction being due purely to good adherence to medication.
Previous research has quantitatively characterised some aspects of medication adherence post-PCI. However, there has not been a detailed exploration of the patient-specific factors relating to such adherence. There is therefore a need to address this deficiency. The aim of this study was firstly to quantify the current level of medication adherence using a validated scale, and then to qualitatively explore the association between the measured adherence and the influencing factors.
Settings and subjects
A convenience sample of 20 patients were recruited to the study. All patients had undergone PCI in the previous 7 days and had completed phase I cardiac rehabilitation. Inclusion criteria included being on three or more cardiac medications (including any of the following: antiplatelets, statins/fibrate/ezetimibe, β-blockers, angiotensin-converting enzyme inhibitors, angiotensin 2 receptor blockers, nitrates, nicorandil, calcium-channel blockers, antiarrhythmics), age of 18 year or more, fluent in English and being able to give informed consent. Patients were excluded from the study if they had cognitive impairment, had known alcohol or illicit drug use, had a physical or psychological disability inhibiting communication, were using a compliance aid (i.e. dosette box) or resided in a nursing, residential or care home.
The sample size for this project was determined by data saturation caused by repeated thematic recurrence in the qualitative semi-structured interviews. Evidence indicated that up to 25 patients would be required to achieve this.[22,23]
Full ethical approval was granted by the North of Scotland Research Ethics Service on the 22nd March 2010. Patients were given an information sheet about the study by cardiology staff who would normally be involved in the care of PCI patients. After a minimum of 24 h to reflect on that information, if they wished to participate in the study a meeting was set up with a researcher (GFR) where further information about the study was given and written informed consent taken before participation in the study.
Pilot study and validation
A pilot study (two patients) was conducted in the penultimate week of April 2010. Both patients met the inclusion and avoided the exclusion criteria for the study. The pilot study was required to check that the methods, procedures and documentation to be used in the study were acceptable to the research participants, and secondly that the methods used would yield data required to answer the research question.
Data collection and analysis
Completion of consent forms, questionnaires and interviews was conducted by a single researcher (GFR) at Raigmore Hospital, Inverness.
Demographic data were collected regarding the medical, social, financial and educational background of each participant; a full medication history was also taken. This enabled descriptive statistics to be used to characterise the sample.
A review of published adherence screening tools was undertaken (Table 1[24–37]). This identified the Tool for Adherence Behaviour Screening (TABS) as the most appropriate questionnaire to provide an accurate, fast and reliable indication of medication adherence in patients with chronic conditions. The reliability and internal consistency, as well as the construct, criterion and incremental validity of TABS, were proven for use in chronic conditions. The completion of TABS involves the patient answering two sets of four questions using a five-point Likert scale. Answers from the first set give an adherence behaviour score (ABS) and answers from the second set give a non-adherence behaviour score (NABS). ABS of less than 19 out of 20 denote a lack of adherence behaviour whereas NABS of more than eight out of 20 can be defined as non-adherence behaviour. Low scores for ABS are suggestive of intentional non-adherence whereas low scores for question 5 (NABS) suggest unintentional non-adherence. It should be noted that although TABS comprises ABS and NABS, cumulative totals of the scores are unable to be given due to the inverse relationship between the scoring for ABS and NABS. Descriptive statistics were used to characterise the sample.
Table 1. Review of suitability of published adherence scores
Eight questions, easy to administer. Reliability and validity have been proven.
Qualitative semi-structured interview
A semi-structured interview was selected as the most appropriate methodology to explore patients' ideas, concerns and expectations about adherence to medication. Interviews were conducted by the corresponding author over a 7-week period from May 2010. The topic guide for semi-structured interviews was adapted from another study of medication adherence in patients with chronic illness. This was reviewed for appropriateness by the team prior to use. All interviews were digitally recorded before being transcribed by cardiology secretarial staff. Once the interviews were transcribed the accuracy of the transcriptions was scrutinised by the corresponding author. Patient confidentiality was maintained by omitting all names and identifiers, and patient approval for the use of direct quotations was obtained. Once the qualitative data had been transcribed the transcripts were loaded into computer-assisted qualitative data analysis software (Atlas.ti version 6.0.1; Atlas.ti GmbH, Berlin, Germany) which expedited analysis and enhanced ‘closeness’ with the data. A thematic framework was developed to code the transcripts. The original coding framework was agreed upon by GFR and SJL. GFR completed the coding and SJL verified the accuracy of each applied code on all transcripts. Initially there was a process of familiarisation by listening/re-listening to the recorded interviews while reading/re-reading the transcripts, which allowed immersion in the data. A process of ‘coding’ was applied to the transcripts and these codes allowed for themes to be identified. The construction of the initial thematic framework was guided by the research aims and objectives and questions introduced to participants from the topic guides. However, the framework analysis could also be used to identify emergent themes expressed during the interviews, offering a unique flexibility to realise themes from outwith the topic guide. Contextual meaning for each quote and code were then indexed before being displayed in a process called charting. Finally, associations between codes and themes were explored during mapping and interpretation. Relevant quotes supporting the framework could then be displayed. Identifiers for individual patients followed each quote and were given as the patient number, the paragraph number in the transcript, sex, age and TABS scores represented as the ABS and NABS. A framework analysis provided a robust technique for the analysis of qualitative data as it facilitates rigorous and transparent data management.[38,39] This analysis was completed in parallel with recruitment until data saturation was determined.
The rationale for choosing TABS has already been discussed. The TABS questionnaire was validated in another chronic-condition cohort, chronic obstructive pulmonary disease, and was shown to be a reliable score for measuring adherence in a population with chronic disease.
Twenty patients (15 male, 5 female) met the study's inclusion/exclusion criteria and consented to take part – there were no refusals to participate in this research. This sample size achieved data saturation: this was the stage at which no new themes were generated. Eight additional interviews were conducted with no new themes emerging to define data saturation.
Data was wide ranging with regard to age, height and weight of the participants. Only five patients (25%) were found to be of a healthy body mass index (20–25 kg/m2); seven (35%) were clinically obese with a body mass index of more than 30 kg/m2. Male patients comprised 75% of the cohort. The majority of the patients were employed (60%) (Tables 2 and 3).
Table 2. Numerical data summary analysis
Median (interquartile range)
Doses per day
Tablets per day
Body mass index (kg/m2)
Table 3. Categorical data summary analysis
Number of patients
Annual income (GB £)
White British, other
Patients were colour-coded according to their TABS scores (Figure 1). Six patients (30%) (patient numbers 001, 004, 005, 014, 017 and 019) were found to have low ABS (<19/20) (Figure 2). Of those six, only two (patients 014 and 019) were also found to have high NABS (>8/20).
The median ABS for this cohort was 19/20, whereas the median NABS was 7/20; both scores were suggestive of good adherence within the cohort (Table 4). The high value of the median ABS and low value of the median NABS indicated a desire in most patients to take their medication.
Table 4. Tool for Adherence Behaviour Screening (TABS) questions and scoring
Never = 1, % (n)
Rarely = 2, % (n)
Sometimes = 3, % (n)
Often = 4, % (n)
Always = 5, % (n)
Median score per question (interquartile range)
Median score for ABS/NABS (interquartile range)
Adherence behaviour score
1. I have strict routines for using my medicines.
2. I keep my medications close to where I need to use them.
3. I ensure I have enough medications so I do not run out.
4. I push myself to follow the instructions of my doctors.
Non-adherence behaviour score
5. I get confused about my medicines.
6. I make changes in the recommended management to suit my lifestyle.
7. I vary my recommended management based on how I am feeling.
8. I put up with my medical problems before taking any action.
The value of Pearson's r exhibited no correlation between the NABS and the ABS. The clustering of patients in the box on the top left of Figure 2 indicated that 70% of patients scored high for ABS and low for NABS, which is suggestive of good adherence.
Qualitative interview analysis
The full thematic analysis can be seen in Figure 3. The main themes that relate to medication adherence can be found in Figure 4. Most of the themes were positively associated with increased medication adherence. However, the role of adverse drug reactions (ADRs) had a significant negative effect, while the community pharmacist role was considered non-significant by the majority of patients.
In general, the cohort (especially those with low ABS and high NABS) had a good knowledge of commonly experienced ADRs due to medication they were prescribed. However, as 50% of patients with either a low ABS or high NABS expressed knowledge of ADRs this implies there may be an association between that knowledge and non-adherence.
When it says in the leaflet that it can cause irreversible muscle damage and may result in hospitalisation, that's enough to focus my mind!
005: (78). Male, 56 years old, ABS 17, NABS 5
I think the β-blockers seem to make me a bit sleepy. I mean that if I said I would phone someone in the evening, I might be asleep and didn't phone, that sort of thing. Other than that it doesn't hamper me.
004: (5). Female, 59 years old, ABS 18, NABS 8
The importance of the difference between the terms compliance and adherence is demonstrable when considering the quotes and TABS scores of patients 004 and 005 above. While the TABS scores indicate the potential for poor adherence the nature of that association can be further explored by considering the reason for the scores. In these instances the knowledge of ADRs may influence a patient's decision as to whether they wish to be or can be adherent; that is, intentional non-adherence as the result of experiencing an ADR.
Thirteen patients discussed the impact that having an understanding of the indication has for adherence. These ideas varied greatly between patients.
After an operation especially [PCI], I think people have got to understand that certain pills do certain things to the body that helps them, but if they are a bit wary of pills then they are not inclined to take them unless it is explained why they are taking them [and] why they are to take them.
002: (157). Female, 70 years old, ABS 20, NABS 7
Another patient (008) with high ABS and low NABS admitted to not understanding what his medication was prescribed for. However, critically, his adherence remained high because he had rationalised the need for additional medication and therefore perceived a health benefit with the medication.
I know that these tablets are being prescribed for a reason and probably the truth is, what each tablet does for the body, I don't really know, but obviously I have had to receive another couple because obviously number 1 for example doesn't do what number 2 and 3 does otherwise I perhaps wouldn't be on a second or a third, but I do understand that I have to take that medicine.
008: (17). Male, 54 years old, ABS 19, NABS 7
Relationship with doctor
There was a higher frequency of quotes for this code than any other. In total 17 patients offered ideas about the doctor–patient relationship. Of the 17 patients, 16 noted good relationships with their general practitioner (GP). Patient 019 (low ABS and high NABS) described a poor working relationship but was still of the belief that a good relationship was desirable.
A number of patients were also of the opinion that if a medication was prescribed for you by a doctor then it should be taken regardless.
Well to me it is common sense. If the doctor says you need it then you need it so you should take it.
009: (133). Male, 64 years old, ABS 19, NABS 4
I am just the sort of guy that would take the medicine anyway I think having had it prescribed, but, I think it is good to have confidence in your doctor and believe that he knows what he is doing and has your best interests at heart and I think that is the case with my doctor.
012: (77). Male, 79 years old, ABS 20, NABS 4
It's prescribed by the doctors and that is it you would still take it. You know you have such faith in the doctors, well I have, I can't speak for everyone else but I do.
013: (70). Female, 62 years old, ABS 19, NABS 6
The results uncovered a lack of understanding of the role of the pharmacist. Patients did not want to undermine the stature of the prescriber. There was also a misconception that for serious ailments pharmacists have no role to play.
. . . I have never seen the pharmacist in that role, they sort of sit behind a shop counter. I know it is a highly trained profession, so why not? Because once they are prescribed, I have already been to the GP.
020: (238). Male, 52 years old, ABS 19, NABS 7
Not if it was to do with the heart.
014: (182). Male, 65 years old, ABS 16, NABS 9
There would appear to be a view among the cohort that aspirin holds less importance than other medication.
. . . if it is something minor like an aspirin or something, I know I have to take the aspirin for my heart but if I missed one it wouldn't bother me so much.
002: (149). Female, 70 years old, ABS 20, NABS 7
I understand the aspirin is important but I don't think in relation to the other pills it is as essential. But I always take it and I always make sure I have it.
002: (153). Female, 70 years old, ABS 20, NABS 7
. . . the aspirin in less important because that is general thinning.
020: (118). Male, 52 years old, ABS 19, NABS 7
Overall 13 patients in the cohort reported having a routine or system for taking their medication. There was a belief among these patients that having a routine improved their adherence.
I have been taking them for 18 years now so it is just a routine now. It is part of my lifestyle.
009: (69). Male, 64 years old, ABS 19, NABS 4
One of the main tips that these patients had was that by keeping medication in the same place (and preferably visible) this acts as a prompt to take medication.
I have another pill which I take prior to my evening meal. In order not to forget that I also have a whisky before my evening meal! . . . I never forget the whisky . . .
012: (21). Male, 79 years old, ABS 20, NABS 4
I forget almost never. Just basically by keeping it in the same area and doing it at the same time.
003: (57). Male, 65 years old, ABS 19, NABS 5
The experience of severe chest pain and the subsequent knowledge that it was a heart attack acted as a motivating factor to many.
If you know the consequences well. . . I don't want to suffer the consequences of going back in [to hospital] with a heart attack or something like that. It probably does frighten you into taking it and don't miss it out.
016: (89). Male, 59 years old, ABS 20, NABS 6
I understand that the medicines are going to be good for me, I understand the importance of keeping a supply and trying not to miss them, I have only missed the odd one in the evening when I had missed one, I ignored it.
005: (71). Male, 56 years old, ABS 17, NABS 5
I will continue to take it but if I don't think it is suiting at all then I normally put it in the back of the drawer and forget about it.
019: (21). Male, 56 years old, ABS 17, NABS 11
While patient 005 stated that he understood the importance of taking his medication he also admitted to missing doses, questioning the motivation he has to remain adherent. As for patient 019, his quote demonstrates explicitly intentional non-adherence. This quote further explains the reasoning behind the low ABS and high NABS.
Understanding of condition
Having an understanding of your heart condition and the drugs used to treat it was highlighted as a fundamental principle. Once a patient has this knowledge it contributes to their adherence. This process was a key step for patient 020 in establishing a method for ensuring no further MIs.
. . . because understanding the medication is part of understanding the condition, I am not just understanding what happened to me but also trying to make sure that it doesn't happen again, so it is important to understand, for the patient, for me to understand why I am on certain drugs.
020: (34). Male, 52 years old, ABS 19, NABS 7
Patients with low ABS or high NABS
One prominent issue noted in patients with low ABS or high NABS was around ADRs. Four out of the six patients mentioned ADRs during the interview. Importantly they were able to discuss the particular types of ADR they might expect from their prescribed medication. Low ABS or high NABS was not associated with baseline characteristics such as education completed, employment and income.
Summary of main findings
High ABS and low NABS, suggestive of good adherence, were found in 70% of the patients in this cohort. Figure 4 depicts themes derived from patient interviews which impacted on the scores expressed. Each theme is dependent on individual patients' specific beliefs, knowledge and understanding of their own condition. However, attaining high ABS or low NABS is not reliant on expression of all the themes. If patients believed strongly in only one or two themes this could be enough to result in a good score.
On the periphery of these themes, and not as central to medication adherence and certainly not as widespread, are other themes such as information sources, understanding of medication and help from a community pharmacist.
There was a misconception among some post-PCI patients about the potential benefits of taking aspirin. Perhaps the ubiquitous nature of aspirin prescribing may have led to some misconceptions about the efficacy of the medication. This is especially concerning when considering the critical role of aspirin in the prevention of post-PCI complications including stent thrombosis. It seemed as though aspirin was not thought by some patients to be as important as other medications.
Limitations of this research
All 20 patients were recruited into the study within 7 days of undergoing PCI. The close chronological proximity of this study to the procedure and the information given during phase I cardiac rehabilitation may make patients, at the time of recruitment into the study, more inclined to take medication. The sustainability of this adherence was not investigated as it was outwith the scope of the research question.
The cohort studied included patients who had undergone PCI electively or following an acute MI. Whether a patient had experienced an MI or they were having PCI electively may have augmented an increase in motivation to take medication. Those patients who had experienced an MI spoke of excruciating pain, as well as fear of subsequent events. The risk of stent thrombosis to patients from non-adherence with post-PCI medication is however the same. Therefore, it is appropriate to be indiscriminate with the selection of a post-PCI cohort.
The qualitative results of the study are based on interviews with patients. It should be noted that quotations are thus based on accounts of events rather than on specific evidence of those events. Also, from a reflexive perspective, all participants in the study knew they were going to be interviewed by a pharmacist about their adherence to medication. Again, these factors may have influenced the study and the responses for participants.
Comparison with the findings from other literature
This was the first study to explore the patient-specific factors associated with medication adherence in a post-PCI cohort. However, patient adherence to the antiplatelet drug clopidogrel has been measured in two studies of post-PCI patients without characterising the reasons for such adherence. Firstly, Spertus et al. reported that one in seven post-MI patients with a stent stopped clopidogrel by 30 days, resulting in a significant increase in mortality over the next 11 months from 0.7 to 7.5% (P < 0.001). No patients in the cohort studied in this research overtly stated the opinion that they would cease clopidogrel, except on the decision of a doctor. Secondly, Ho et al. reported that discontinuation of clopidogrel increases risk of mortality in post-ACS patients with a stent from 6.9 to 19.9% (P < 0.001). The risk of not being adherent with the post-PCI antiplatelet regimen is evidently potentially life-threatening. In light of the discovery in this research, greater emphasis should be placed on the importance of aspirin, both by the healthcare professional and for the patient by means of appropriate education about the risks of death.
The proportion of patients with high ABS and low NABS, suggestive of good adherence, was considerably higher than the 50% mean adherence rate for patients on medication for long-term conditions. The results presented give an insight into patient-specific themes relating to adherence behaviour as well as quantifying that behaviour.
Implications for practice
For some patients the role of the community pharmacist was not well understood. The implementation of new and advanced pharmaceutical services for patients with long-term conditions, like the Chronic Medication Service (CMS) in Scotland, will increase the profile of the community pharmacist among other healthcare professionals in the management of chronic disease states. It remains to be seen what impact this new role for community pharmacists will have on increasing adherence in patients. However, as this research has shown, it is imperative that patients have a good relationship with their doctor, or other healthcare provider if this role is devolved.
By delivering personalised care (a tailored approach to medication prescribing and practice) specific needs of individual patients can be met. Personalised care would draw from information, advice, support, feedback and continued education based on the themes identified in this research to provoke and invoke adherence. Only then can the prescriber–patient relationship attain good adherence though concordance. This involves migration away from the historical paternalistic prescriber-led consultations to one in which the patient feels they have a key role to play. Principally, the issue here is one of prescriber cognisance while prescribing.
The results are suggestive of an association between patients' beliefs, knowledge, understanding and misconceptions about medication and their adherence. The nature of such an association is dependent on themes relating to prescribed medication, communication and information, disease, individual patient factors and in particular misconceptions about all the above. However, the associations between the specific themes relating to adherence and an individual patient's adherence are not uniform. They are instead individual, pertaining exclusively to each patient. Increasing adherence therefore has to be tailored to the needs of the individual. Interventions should draw upon the themes relating to adherence outlined in this research, before selecting the most appropriate course to meet the needs of the individual. Essential to the understanding of the themes required is an understanding of the patient by the healthcare team and in particular the prescriber.
Conflict of interest
The Author(s) declare(s) that they have no conflicts of interest to disclose.
This research was supported by the NHS Highland Research & Development Committee Endowment Fund.
The authors would like to sincerely thank the research participants for their participation in this study. We are grateful to the staff of Raigmore Hospital, Inverness, for their time and cooperation during the recruitment phase of this project. The authors would also like to acknowledge Dr Johnson George for the use of the TABS score in this study.