METHODS

Study participants

Patients presenting a prescription in one of 23 community pharmacies in south east England were recruited opportunistically between March 1999 and February 2000. They were eligible for inclusion in the study if they were picking up a new medication for a chronic condition and were aged 75 years or older or had one or more of the following chronic conditions: stroke, coronary heart disease, asthma, diabetes, and rheumatoid arthritis. These conditions were chosen because they are specific priorities for the NHS. Exclusion criteria were the inability to understand spoken English, no telephone, or aged under 18 years. The pharmacists who were recruiting patients received training beforehand.

Written informed consent was obtained from all participants and the study was approved by the Northern multicentre research ethics committee and local research ethics committees.

Measurement tools

Patients received a telephone interview from a researcher 10 days after recruitment; those who responded received another at 4 weeks. A semi-structured telephone interview, which had been piloted and revised, was used to obtain information to meet the study objectives. We asked patients about their adherence, problems, and “issues” with their medication and their information needs. Open ended questions were used and patients’ responses were recorded verbatim where possible. A postal questionnaire was sent immediately after each interview to establish demographic information and health status using the question on general health from the SF36 questionnaire.⁸

Measurement of adherence

Patients’ self-reports about their non-adherence were chosen since those who report non-adherence tell the truth⁹ and this method detects problematic levels of non-adherence.¹⁰ Non-adherence was determined from the question:

“People often miss taking doses of their medicines for a whole range of reasons. Thinking first of the medicine that you started taking when you started on this study (that is, .........), when was the last time you missed taking a dose of this medicine?”.

Patients were defined as non-adherent if they had missed any doses in the previous 7 days; this was followed up by probes for an account of the reason. We chose this definition as it has been suggested that low adherence is indicated if one or more doses are missed.¹⁰ Furthermore, theories of human error suggest that all deviations from correct action are of interest.¹¹ The reason given for non-adherence was categorised as either intentional or unintentional by two of the researchers separately (SC and NB); any discrepancies were resolved by discussion. We also distinguished between partial and complete non-adherence. Partial non-adherence was defined as missing one or more doses of medication and complete non-adherence as completely stopping the medication without consulting the GP.

Analysis of data

Qualitative data from the open ended questions were coded by researchers (JP, SC) and validated by a pharmacist (NB); any mismatches were discussed and agreed upon. SPSS version 10 was used to store all the data and to analyse the quantitative data.

RESULTS

A total of 258 patients were recruited from 23 community pharmacies, of which 239 were interviewed at 10 days and 197 at 4 weeks (fig 1). The characteristics of the patients are shown in table 1.

Adherence

Of the 239 patients interviewed at 10 days, 13 had stopped taking the medication on medical advice. Table 2 shows the extent of non-adherence at 10 days and 4 weeks. Of the 226 remaining, 67 patients (30%) were non-adherent to their new medication. There were 26 cases of non-adherence to other prescribed medication, eight of which were patients who were also non-adherent to their new medication; the remaining 18 had been completely adherent to the new medication. Overall, 85 patients (38%) were non-adherent to at least one of their prescribed medications. At 4 weeks 197 patients were interviewed, 26 of whom had stopped their new medication on medical advice. Of the 171 remaining, 43 (25%) were non-adherent to their new medication; 19 of these had also been non-adherent at 10 days and 24 had previously been adherent.

The proportion of intentional to unintentional non-adherers was similar at 10 days (45% v 55%) and 4 weeks (44% v 56%). None of the unintentional adherers had completely stopped taking their medication but, of the intentional non-adherers, 18/30 (60%) had completely stopped taking their medication at 10 days and 17/19 (89%) at 4 weeks.

Problems caused by medicines

At the 10 day interview 208 patients were still taking their new medication (of the 239, 13 had stopped on medical advice and 18 were completely non-adherent). Of these 208, 138 (66%) reported at least one problem or issue with their new medication. In total, 241 problems/issues were reported and were coded into one of three categories: side effects (n = 121, 50%), concerns about the medication (n = 103, 43%), and difficulties with the practical aspects of taking the medication (n = 17, 7%). Examples are shown in table 3.

Of the 138 patients reporting a problem/issue at 10 days, 56 (41%) reported at 4 weeks that their problem had been resolved but 31 (22%) reported that they were still experiencing a problem; data were unavailable for 51 (37%). Thirty six patients (26%) reported new problems emerging at 4 weeks; 12 of these had not reported a problem at 10 days. Most of the problems were related to side effects (n = 23, 64%), eight (22%) were concerns about the medication, and five (14%) were related to practical aspects of medication taking.

Information needs

There was a substantial and sustained need for further information expressed in the interview. At 10 days 137 of the 226 patients still taking their medication (61%) said they had further information needs, as had 88 of 171 (51%) of those still taking their medication at 4 weeks. These needs were consistent at both time points with about half wanting more information on their medication (n = 59 (43%) at 10 days; n = 46 (52%) at 4 weeks) and over a quarter wanting information on their condition (n = 38 (28%) at 10 days; n = 26 (30%) at 4 weeks).

Of the three areas of concern (non-adherence, problems, and information needs), only 37 (16%) of the 226 patients still taking their medicine at 10 days said they were adherent, problem free, and had received sufficient information. One area of concern was present in 86 (38%) of cases, two areas in 65 (29%), and three areas in 38 (17%). Of the 67 non-adherent patients, 38 (57%) additionally said they had problems and information needs. A greater proportion of intentional non-adherers declared problems with medicines (26/30, 87%) than unintentional non-adherers (22/37, 59%, p = 0.016, Fisher’s exact test) at 10 days.

DISCUSSION

Patients who started a new medication for a chronic condition experienced considerable problems. About one third did not take their new medication as prescribed, almost half of these deliberately so. The incidence of non-adherence was greater with new than with existing medication. Patients frequently encountered problems and had substantial unmet needs for information and support. While some problems were resolved, many were not and new ones were reported.

This rapid emergence of non-adherence and the related plethora of problems have not previously been recognised, nor has the fact that, although some problems are solved, new ones emerge in the first month. Why is this the case? Two factors seem likely. The first is that the current prescribing and dispensing processes are of limited effectiveness. The second is that patients have needs that can only emerge once they have experienced taking a medication.

There is an extensive literature that chronicles the inadequacies of the current processes. Doctors give little information about medication in their consultation and often make inappropriate judgements about the expectations of their patients.¹² The style of communication used by doctors for patients with chronic conditions is one that concentrates on the condition rather than the whole person, which results in less satisfaction and more non-adherence.¹³ Even when patients are given information, they often misunderstand what their doctor says and fail to recall much of the information they are given.¹⁴ Pharmacists have the potential to rectify many of these problems when dispensing medication but often do not; 58–90% of patients report never or rarely receiving unsolicited advice from a pharmacist.^15,16

Even when information about a medicine is given to a patient, it is not specific to that patient but describes a population response to the medication. However, once the patient experiences the medication they get some knowledge of what it does to them and new questions will arise. Patients will form beliefs about the necessity of the medication and formulate concerns about the consequences of its use—factors which have been shown to influence adherence.¹⁷ They will want to know whether it is working or not and, if they feel different in some unpleasant way, whether the medication caused it or not. It is at this point, when patients need new questions and concerns to be resolved, that they are failed by existing medical services. Given their information gap, they consult with friends and relatives and try to reason their way out using their existing beliefs about their condition, their medications, and the way they work. This process can involve false beliefs or an imbalance of correct beliefs about the medication, all of which may encourage inappropriate non-adherence.

The main limitations of this study are the serendipitous nature of recruitment and the use of self-reports of non-adherence. The opportunistic sampling, while geographically widespread, may not be representative of the population as a whole. The use of self-reported non-adherence may under-report the true incidence of non-adherence; if this is the case, the results provide even greater cause for concern.

The findings of this study could be used to contribute to improving the quality of patient care. It seems there is a role for a new service to support and advise patients in their early days of medication taking and to change prescribing decisions if necessary. The NHS has ambitious new plans for pharmacists¹⁸ that include management of repeat prescribing and wider prescribing powers. It may be that, in collaboration with prescribers, new pharmacy services can be developed to meet this substantial need.

Note that our view of good practice is not that adherence is always right. We are taking non-adherence as a marker that there is a problem in the process of prescribing and the experience of using medicines. It is likely that these findings have significance beyond UK practice as the incidence of non-adherence is similar worldwide.^1,4,5