Lithium is an effective therapeutic agent in the control of mood in affective disorders (Reference Baastrup and SchouBaastrup & Schou, 1967; Reference Prien, Caffey and KlettPrien et al, 1973; Reference Prien and PotterPrien & Potter, 1990). However, the dramatic success of lithium has been tempered by poor levels of patient compliance (Reference Van PUTTENVan Putten, 1975; Reference Bech, Vendsborg and RaphaelsenBech et al, 1976; Cochran, Reference Cochran1984, Reference Cochran1986). Various factors have been associated with patient compliance in affective disorders (Reference Jamison, Gerner and GoodwinJamison et al, 1979; Reference Connelly, Davenport and NurnbergerConnelly et al, 1982). Harvey and Peet (Reference Harvey and Peet1991) found that a controlled education programme resulted in improved reported tablet omissions and serum lithium levels, suggesting that knowledge and attitudes were related to compliance. Patient satisfaction has been claimed to increase the likelihood of medication compliance in a range of medical settings (Reference DavisDavis, 1968; Reference Francis, Korsch and MorrisFrancis et al, 1969; Reference LeyLey, 1976; Reference Ley, Bradshaw and KinceyLey et al, 1982). However, little is known about the role of satisfaction in compliance in patients suffering from affective disorder.

Lithium non-compliance is a prevalent problem in patients with affective disorder (Reference Strober, Morrell and LampertStrober et al, 1990). The current study aims to investigate the role of the following factors in compliance: patient satisfaction, lithium attitudes and knowledge and prescribed dosage of lithium. Compliance will be examined using a wide range of measures. By evaluating the possible precursors of lithium compliance, this might bring us closer to identifying and managing lithium non-compliance more effectively.

The study

Patients eligible for study entry had a clinical diagnosis of affective illness, were in attendance at a lithium maintenance clinic and had been prescribed with lithium or carbamazepine. Patients with schizoaffective disorder were excluded from the study. A total of 42 patients were approached. Thirty patients gave informed consent (28 lithium, 2 carbamazepine) and comprised 14 women and 16 men whose mean age was 45.3 years (s.d.=12.4; range 25-64 years). Eighteen patients were clinically diagnosed with manic depression and 12 with major depression. The mean age of onset of the affective disorder was 30.4 years (s.d.=12.6; range, 13-60 years). Patients had been taking lithium for an average of 7.7 years (s.d.=9.4; range, 5 months—30 years). Thirty per cent of patients had received formal psychological or psychiatric counselling of significant duration. Twenty-four patients were receiving additional medication (range, 1-9). Data were not available for the 12 patients who declined to give consent to participate.

A staff psychiatrist saw all patients at the lithium maintenance clinic and invited those willing to participate to sign a patient consent form. Ethical approval from the local health board was granted for the study. Questionnaires were administered by an independent assessor during a single appointment, under the supervision of a research clinical psychologist. The following measures were taken.

Findings

Details of the results of the compliance assessment are shown in Table 1.

Table 1.

Compliance assessment

Compliance rating	Mean score
Patient self-report	‘very good’
Clinician report	‘very good’
Serum lithium level	0.66 mmol/l ‘acceptable’
Previous acceptable assays over past 12 months	94.6%
Number of missed appointments over past 12 months	0.4 (75% of those who missed appointments were manic depressed)
History of any non-compliance?	seven patients with manic depression (23%)

Comment

The patients investigated in the present study were all defined as currently compliant. This finding contrasts markedly to previous research, where approximately one-quarter of patients maintained on lithium regimens failed to comply with physicians' instructions (Reference Jamison, Gerner and GoodwinJamison et al, 1979; Reference Connelly, Davenport and NurnbergerConnelly et al, 1982; Reference CochranCochran, 1986). It should be noted that the patients in the current study were all attending a specialist lithium maintenance clinic and this may account for the high compliance rates found here. The lack of association obtained between compliance and diagnostic category is consistent with the findings of Connelly et al (Reference Connelly, Davenport and Nurnberger1982). However, patients who had a history of non-compliance had manic depression, were mostly male and tended to rate themselves as less compliant than those with no history of non-compliance. Thus, poor previous adherence to lithium regimens is generally predictive of poor current levels of compliance, supporting the trends observed by Frank et al (Reference Frank, Prien and Kupfer1985). Patients with a history of non-compliance had been prescribed higher doses of lithium, possibly leading to the increased presence of side-effects (Reference Jamison and AkiskalJamison & Akiskal, 1983; Reference Gitlin, Cochran and JamisonGitlin et al, 1989). Also, more unstable illness may have produced both poorer compliance, causing more symptoms, and a higher subsequent prescribed dosage of lithium. Patients perceived the standard of care received at the clinic to be very high. The weak and non-significant association between patient satisfaction and compliance was inconsistent with the strong association found by Francis et al (Reference Francis, Korsch and Morris1969). In the present study, however, any possible association may have been masked owing to the high ceiling effect for both variables. Also, our sample was small and may not have been representative of all lithium patients, as those who did not participate may have been dissatisfied. Patient knowledge about lithium was moderate and decreased with age. Attitudes toward lithium were generally poor, particularly among patients with manic depression and with increased years taking lithium. In contrast to the findings of Peet and Harvey (Reference Peet and Harvey1991) LKT and LAQ scores were not associated with compliance. The LKT scores were similar to the community psychiatric nurse scores obtained by Peet and Harvey (Reference Peet and Harvey1991), and also similar to the scores of Peet and Harvey's (Reference Peet and Harvey1991) patients after their educational programme.

Cade's (Reference Cade1949) original claim that lithium could be described as the 'magic wand' for patients can only be supported in the presence of adequate current compliance. Patients in the current study were all currently compliant, in striking contrast to previous research. These findings collectively have clinical relevance for existing lithium clinics, and it is suggested that educational interventions should be directed particularly towards older patients with manic depression who have been taking lithium for a number of years.