Portal de Periódicos da CAPES

Adherence improvement in Dutch glaucoma patients: a randomized controlled trial

2012; Wiley; Volume: 91; Issue: 7 Linguagem: Inglês

10.1111/j.1755-3768.2012.02571.x

1755-3768

Henny J. M. Beckers, Carroll A.B. Webers, Michiel J. W. M. Busch, Hendrik M. A. Brink, Thomas P. Colen, Jan Schouten,

Ocular Surface and Contact Lens

Purpose: To study the effect of patient education and the TravAlert®-Eyot® drop guider on intraocular pressure (IOP) and adherence in patients with primary open angle glaucoma (POAG) or ocular hypertension (OHT) monitored with the TravAlert® dosing aid. Methods: Multicentre, randomized, controlled clinical trial among 18 Dutch hospitals. Patients were randomized to one of the four study arms: (1) use of the dosing aid, (2) use of the dosing aid with the drop guider, (3) use of the dosing aid together with patient education or (4) use of the dosing aid and drop guider together with patient education. IOP was recorded at baseline and after 3 and 6 months. Data on adherence generated by the dosing aid were collected and studied at the end of the study. Results: Mean IOP dropped from 20.3 ± 5.7 mmHg at baseline to 16.3 ± 4.0 mmHg (right eye) after 6 months and from 20.2 ± 5.9 mmHg to 16.4 ± 4.1 mmHg (left eye). The mean adherence rate was 0.91 ± 0.1. IOP and adherence rate were not statistically different between the study arms. Patients with 'drug holidays' had a significantly higher mean IOP after 6 months. Patients who used the drop guider were less adherent. A lower adherence level was also associated with new patients with glaucoma and patients with a lower level of knowledge on glaucoma. Conclusion: Patient education is especially useful for new patients with glaucoma. The use of a drop guider does not improve adherence. Especially patients with 'drug holidays' are at risk for developing uncontrolled IOP levels. Nonadherence with prescribed topical therapy is a frequently encountered problem in patients with primary open angle glaucoma (POAG) or ocular hypertension (OHT). In a recent review, the proportion of patients not fully following their treatment as prescribed was found to vary from 4.6% to 80%, because of varying definitions of noncompliance and different assessment methods (Olthoff et al. 2005). Progression of the disease, and thus visual field loss, may be caused by nonadherence (Stewart et al. 1993; Collaborative Normal-Tension Glaucoma Study Group 1998). Ophthalmologists should be aware of the possibility of nonadherence in cases when prescribed topical therapy is not effective. For medications to be effective, patients need to take them (Cramer 2002; Schwartz & Quigley 2008). Especially patients with glaucoma are at risk to become less motivated to use their medication as prescribed and thus may become nonadherent, because of the lack of symptoms in early stages of the disease (DiMatteo et al. 2002; Friedman et al. 2007b). Unfortunately, nonadherence is often not detected in daily practice because patients conceal it from their doctor, possibly because they want to be perceived as 'good patients' (Kavookjian et al. 2005). However, the magnitude of the impact of nonadherence on intraocular pressure (IOP) levels and visual field loss has still not been clearly elucidated. In one study where a questionnaire was employed, nonadherent patients with glaucoma were found to have higher IOP levels and more severe visual field loss (Konstas et al. 2000). A recent study that distinguished patient groups with high, medium or low adherence levels reported higher mean IOP in patients in the low adherence group (Nordmann et al. 2010). Another study reported more blindness among nonadherent patients (Forsman et al. 2007). A frequently observed problem in patients with glaucoma is the inability to correctly administer the eye drops, including difficulty in aiming the drop, squeezing the bottle and seeing the tip of the bottle (Winfield et al. 1990; Kholdebarin et al. 2008). Patients often rely on partners or relatives to put in the drops for them. Correct instillation of eye drops by patients with glaucoma themselves could therefore be beneficial to improve adherence. The intervention studies that have been undertaken mainly aimed at reducing forgetfulness (Olthoff et al. 2005). A recent review showed that interventions involving simplified dosing regimes, reminder devices, education and individualized care planning did show improvements in adherence rates (Gray et al. 2009). However, from these data, no particular interventions could be advised. It is still unclear to what extent patient education by improving knowledge may help to achieve better adherence with medication. A beneficial effect of patient education has often been presumed (Olthoff et al. 2005; Budenz 2009). In recent studies, using a knowledge questionnaire, it has been shown that an individual approach using interventions focused on attitude and discipline may be the best approach to improve adherence. People with a lower education level may benefit the most from information on public assistance and practical aspects of glaucoma (Hoevenaars et al. 2006, 2008). The TravAlert® dosing aid is a monitoring device for the use of travoprost 0.004% (Alcon Laboratories, Inc., Forth Worth, TX, USA) or the fixed combination of timolol 0.5%-travoprost 0.004% (Alcon Laboratories, Inc.) eye drops. It contains an easy-to-squeeze lever and can be used with visual and audible alerts. It collects data on the number of drops administered on specific days using a drop log calendar and a dosing log. The use of an additional drop guider (TravAlert®-Eyot®) may also facilitate the correct instillation of eye drops. The dosing aid has been shown to accurately record the administration of eye drops (Boden et al. 2006; Cronin et al. 2007; Friedman et al. 2007a). Our objective was to study the effect of patient education – assessed by a knowledge questionnaire –, the use of the TravAlert®-Eyot® drop guider and their combined effects on IOP and adherence in POAG or OHT patients who used the TravAlert® dosing aid and started with (or added or switched treatment to) travoprost 0.004% or the fixed combination of timolol 0.5%-travoprost 0.004%. Additionally, patient satisfaction with the dosing device and drop guider was studied. A randomized clinical trial with a 2 × 2 factorial design and a duration of 6 months after inclusion was performed in 18 Dutch participating hospitals. Institutional Review Board (IRB)/Ethics Committee approval was obtained from all participating hospitals before the study was started. The study was conducted according to the principles of the Declaration of Helsinki (version 2004) and in accordance with the Medical Research Involving Human Subjects Act. It was registered at ClinicalTrials.gov (identifier NCT00603005). All outpatients with a diagnosis of POAG or OHT and a minimum age of 18 years who were either starting treatment with travoprost 0.004% or the fixed combination timolol 0.5%-travoprost 0.004%, switching from previous medication to travoprost 0.004% or timolol 0.5%-travoprost 0.004% (because of side-effects or insufficient IOP lowering), or adding travoprost 0.004% or timolol 0.5%-travoprost 0.004% to previous medication (because of insufficient IOP lowering) were asked to participate. However, patients who were expected to have a high risk of side-effects from travoprost or the fixed combination travoprost/timolol, based upon their medical history and/or previous use of similar medication, were excluded. Other eye medication and/or systemic medication were allowed. All patients who agreed to participate signed an informed consent, after which they were randomized to one of the four study arms: Use of the TravAlert® dosing aid, Use of the dosing aid with the TravAlert®-Eyot® drop guider, Use of the dosing aid together with patient education and Use of the dosing aid and drop guider together with patient education. Education on glaucoma and its treatment – the need to adhere to the prescribed treatment included – was given by research assistants, who were trained by the authors to use a standardized protocol to provide identical information to all study participants from all participating centres. To reach this goal, two authors who are experienced glaucoma specialists (HB and CW) developed a 15-min instructional video in which the main aspects of the disease and its treatment were shown to the participating patients in a comprehensive fashion. This video was approved by the other authors and research assistants. At baseline, the following patient characteristics were recorded: gender, age, educational level, time since glaucoma diagnosis and amount of visual field loss. To determine visual field loss, one recent reliable visual field (max. 6 months old) was requested for all study patients. IOP levels were recorded as part of a routine eye examination. Study visits were scheduled at 3 and 6 months. During these visits, IOP was measured again. A research assistant collected the data from the dosing aid and helped patients with questions on the use of the device. Additionally, at 3 months, patients who previously received patient education on glaucoma were again shown the standardized, instructional video. During the study, the investigators were masked for the intervention (drop guider/patient education) their patients were receiving. In return, the participating patients were encouraged not to report the intervention they were receiving. To determine the level of knowledge on glaucoma and glaucoma treatment, a questionnaire was used at baseline and at the end of the study, which contained 19 items on knowledge of glaucoma and 18 items on knowledge of glaucoma treatment. This questionnaire had previously been used (Hoevenaars et al. 2005, 2006, 2008) and had been developed in a systematic way, based on focus group interviews, expert opinions and a pilot test (Hoevenaars et al. 2005). At the end of the study, the participants were asked their opinion on the use of the dosing aid (and drop guider, if applicable). Also, they were asked to report how many days that they did not take their drops during the past 4 weeks. Owing to the lack of customization of TravAlert® software in the drop log calendar and dosing log, and the fact that participating patients often changed their planned dosing time point according to their own convenience, the dosing aid was only used to capture the time points of drops taken by the patients. These time points were collected, transferred and then analysed with self-developed software. A dose was counted as taken on a day if the number of drops registered by the dosing aid was greater than zero before 4 am on the next day. Apparently, some patients took too many drops per day. They might have pushed the lever accidentally or might have taken too many drops on purpose. Because we could not verify how many drops actually reached the eye, we regarded the dose as taken if at least one drop was registered by the TravAlert® device. Some technical problems of the dosing aid, such as running out of battery power that resulted in resetting the internal clock to 1-JAN-1980, were corrected in the analysing software. To correct the problem with the internal clock, we used the following algorithm: 'If drop date is before start date AND date difference is negative THEN corrected drop date is drop date plus the absolute value of date difference UNTIL corrected drop date is after end date'. This should result in an ascending order of the drop dates. Furthermore, the start date was corrected when it was not identical to the visit day. The start date was defined as the day the participant started using the dosing aid. For the end date of the analysis, the last registered dosing date within 7 days before the closing visit date was used. The adherence rate was calculated as the ratio of the recorded number of adherent days to the total number of study days. Because the device can make extra recordings when the lever is depressed erroneously, we included a maximum of 1 dose taken per day in the calculation. Several different adherence patterns were distinguished and categorized as follows: 'Adherent' (a mean of 0–1 missed doses per month over a 6-month period), 'medium adherent' (a mean of 2–6 missed doses per month over a 6-month period), 'nonadherent' (a mean of 7 or more missed doses per month over a 6-month period) and 'drug holiday'. Subjects were included in the drug holiday group if the number of consecutive no-dosing-days were greater than or equal to 8 (according to the definition by Ajit et al. 2010). The statistical analysis was performed using SAS software version 9.1.3 (SAS Institute, NC, USA). The power was 90% (α = 0.05, two tail). Analysis of variance (anova) was used to compare continuous data, and categorical data were analysed using the Pearson chi-square test. Student's t-test was used to compare the adherence pattern of drug holiday patients versus other adherence patterns. General linear model was used in a multivariable analysis to find any associations between adherence rate and other parameters. IOP levels were calculated separately for each eye of the study participants. The scores of the questionnaire on knowledge of glaucoma and its treatment – possible answers 'correct', 'incorrect' or 'don't know'– were calculated as the sum of correct answers. Between October 2007 and June 2009, 805 patients gave their consent and were enrolled in the study (Fig. 1). From these, 652 patients successfully completed the study, while 153 patients dropped out prematurely. The reasons given for dropping out were: advice by the ophthalmologist to stop the prescribed medication – mainly because of the occurrence of side-effects – (n = 59), the participation in the study was too bothersome or the patient could not handle or would not use the dosing device (n = 50), becoming ill or admittance to hospital (n = 13), other reasons (e.g. removal to another city, long holiday, private circumstances, n = 16) and no reason given (n = 6). Additionally, 3 patients died during the study, and 6 patients were lost to follow-up. The participants who dropped out for the reason that they found participation too bothersome or could/would not use the dosing device were significantly older (often 80 years or older). Furthermore, significantly more women dropped out for these reasons. Flowchart of study participants. Owing to technical problems, such as running out of battery power, the device did not correctly store, or only partly stored, the data in a number of patients. Therefore, complete TravAlert® data could be retrieved only from 588 of the 652 patients. Analysis was thus limited to the 588 patients with complete data. The baseline characteristics of the patients in the 4 study arms were found to be comparable (Table 1). There were no statistical differences in gender, age, amount of visual field loss, time since glaucoma diagnosis, educational level, level of knowledge on glaucoma and its treatment or IOP levels. The mean age of the participants was 66.3 ± 10.6 years, ranging from 23 to 92 years, and 54% were men. Mean duration of POAG or OHT after diagnosis was 4.9 ± 6.6 years. At the start of the study, 42.2% of the participants were new on glaucoma medication. Visual field loss as determined by the Humphrey Field analyser (Carl Zeiss Meditec, Dublin, CA, USA), showed a mean MD in the right eye (RE) of −4.9 ± 6.4 dB and a mean MD in the left eye (LE) of −4.6 ± 6.2 dB. However, 22% of the study patients had a loss of more than 10 dB in their worst eye. Twenty-seven per cent of the participants had a higher education level, that is, at least a high school diploma or a university degree. Mean IOP dropped from 20.3 ± 5.7 at baseline to 16.3 ± 4.0 RE after 6 months and from 20.2 ± 5.9 to 16.4 ± 4.1 mmHg LE (Table 2). These differences were not statistically significant between the four study arms. The overall mean adherence rate over 6 months of follow-up was 0.91 ± 0.1. Thus, taking all patients into account, 91% of the daily doses were taken over a 6-month period. The differences in adherence rate between the study arms were also not statistically significant (Table 2). However, there was a statistical significant difference between adherence patterns within the four study arms (p = 0.0056, Table 2). More adherent patients – adherence rate ranging from 0.97 to 1.0 – were found in study arm 4: participants receiving patient education on glaucoma. More nonadherent/drug holiday patients – adherence rate below 0.8 or at least 8 consecutive missed doses – were found in study arm 2: participants using the drop guider. Even though a larger number of adherent patients were found in the arms with patient education, the difference in adherence between study participants who did or did not receive patient education was not statistically significant (Table 3). However, there was a statistical significant difference between patients who used the drop guider and those who did not (Table 3): patients who used the drop guider were less adherent. Additionally, patients with drug holidays had a significantly higher mean IOP in comparison with the other adherence patterns (Table 4). Table 5 shows the patient characteristics of the different adherence patterns. There were no significant statistical differences in gender, age, education level, degree of visual field loss and patient education between the different patterns (Table 5). However, a lower adherence level was associated with a shorter time since glaucoma diagnosis (p = 0.01) and with treatment naïve patients (p = 0.03). In Table 6, it is shown that a lower adherence level was also associated with a lower mean score on items concerning knowledge of glaucoma (p = 0.003) and its treatment (p = 0.001) of the questionnaire at the end of the study. These univariate results match the multivariable results to a certain extent: knowledge of glaucoma, knowledge of glaucoma treatment and patients new on glaucoma medication were to some degree associated with the adherence rate/adherence patterns. However, the R2 of the multiple regression analysis was close to zero (<0.07), which indicates that there is only a small response to the variation of its regressors. The results from the questionnaire at the end of the study revealed statistically significant higher mean scores on knowledge of glaucoma (p < 0.0001) and its treatment (p = 0.0009) in the study arms with patients receiving additional patient education (Table 7). Patients who were new on glaucoma medication had a significantly lower knowledge level of glaucoma and its treatment at baseline (Table 8). Generally, patients reported to be satisfied with their dosing device. At the end of the study, 89% of patients who solely used the dosing aid and 79% of patients who used the drop guider together with the dosing aid reported to be either satisfied or even very satisfied. At the end of the study, self-reported adherence by the study participants did not correspond with the degree of adherence registered by monitoring with the dosing aid, as 67% of patients with a medium or worse adherence pattern claimed to have only missed 0–1 doses over the past 4 weeks. These results show a high adherence level of patients who participated in this relatively large Dutch study. This high level of adherence might be better than in real life, because the patients who agreed to participate in the study may have been highly motivated. Furthermore, the participants also knew that they were monitored and might therefore have been eager to obtain good results. However, a significant number of patients dropped out of the study, which may have biased our results. Where adherence data were collected (in 50% of cases), we also found that these patients were less adherent. Those who dropped out because they found participation too bothersome were more often women and older, often even 80 or older. In earlier studies, it has been shown that patients who are younger than 50 or older than 80 are more likely to be less adherent (Friedman et al. 2009; Hermann et al. 2011). The differences between the four study arms with regard to 6-month IOP levels were not statistically significant. An explanation for this finding could be that the adherence rates for all four arms were high, thus probably masking higher IOP levels for the patients with a lower adherence level. Additionally, there might have been no significant effect on adherence by the interventions. Even though we found that patient education led to a higher level of understanding of glaucoma and its treatment, it could unfortunately not be proven that improving knowledge leads to a better adherence level. A drawback of determining the mean adherence rate over 6 months was that 'drug holidays', during which drops were not taken for several consecutive days, could not be distinguished from other forms of nonadherence. Categorizing the level of adherence of the participating patients into patterns, described as 'adherent', 'medium adherent', 'nonadherent' and 'drug holiday', revealed a statistically significantly higher IOP in the 'drug holiday' group at the end of the study. In a recent UK study in which also patterns of adherence were described, no differences in IOP were found between adherent and nonadherent patients (Ajit et al. 2010). Contrastingly, another study from France found higher IOP in patients with the lowest adherence (Nordmann et al. 2010). Unfortunately, the latter study did not distinguish a drug holiday group. Perhaps the number of participants in the UK study was too low to pick up a difference in IOP. Furthermore, also no association between IOP and adherence was found in a study carried out in the USA (Okeke et al. 2009a,b). An explanation offered in that particular study was that patients tended to improve their drop taking shortly before an office visit. In our study, it appeared that a large number of participants (43%) who took drug holidays did so in the last week before a study visit. This may explain why we were able to detect an effect of drug holidays on IOP where other studies could not. Owing to higher IOP levels, especially patients with a 'drug holiday' adherence pattern might be at risk for glaucomatous progression on the longer term. Identifying these patients at an early stage of their disease would therefore seem to be beneficial to improve treatment results. We found that patients were generally satisfied or even very satisfied with their dosing aid. Nevertheless, the results of study arms 2 and 4 suggest that the drop guider was not found to be very useful, and the patients who used the drop guider were less adherent and generally less satisfied with their device than the other participants. The participants had to extend their neck far backwards to use the drop guider on top of the dosing aid. This might have been a problem for elderly patients. Overall, we cannot rule out that in this study, even with the dosing device, patients may have asked others (e.g. their partners) to administer the drops for them. This was technically possible in the study arms in which no drop guider was used, but not possible when the drop guider had to be used. With the drop guider on top of the dosing aid, patients were obliged to put in the drops themselves. Another possibility may have been that the patients sometimes took their medication directly from the bottle, without using their dosing device. In that case, they may still have been adherent but their drops were not registered, leading to a worse adherence pattern in our data collection. In confirmation with earlier data, we again found that patients tend to overestimate their medication use (Jampel et al. 2003; Flowers et al. 2006; Okeke et al. 2009a; Ajit et al. 2010). At the end of the study, a majority of patients stated to be adherent while the monitoring showed them to be not. Patients may not always remember correctly whether they have or not have taken their medication. This was why we asked patients to report their adherence over a period of 4 weeks instead of over the total follow-up period of 6 months. However, as suggested earlier (Okeke et al. 2009a), some patients may also improve their drop taking in the weeks leading up to their study visit. There is also some evidence in literature that patients wilfully try to hide their nonadherence from their doctor, because they want to be perceived as 'good patients', even if they know that they are monitored (Hahn 2009). In an earlier study in which monitor recordings with patient log books were compared, it was found that patients did not report their values correctly, or even made them up (Mazze et al. 1984). On the other hand, a recent study showed that the outcome of adherence measurements of patients who are aware or who are not aware of being monitored might be similar (Hermann et al. 2011). It remains unclear to what extent patient education may help to achieve better adherence with medication. We found that patient education led to a higher level of knowledge on glaucoma and its treatment. Although a better adherence with glaucoma medication in the study arms of patients who received patient education was observed, statistical significance was not reached. Perhaps the duration of the study was too short to reach a substantial effect on adherence, but it may also be possible that the effect of improving knowledge on adherence was small. Thus, the impact of patient education on adherence by only improving knowledge and understanding of the disease may probably be disappointing. It would be interesting to investigate other forms of patient education, such as interventions focused on attitude and discipline, as suggested earlier (Hoevenaars et al. 2006, 2008). A lower adherence level was also found to be associated with a shorter time since glaucoma diagnosis and with patients who were new on glaucoma medication. The latter group also showed a lower level of knowledge at baseline. These findings indicate that patient education by improving knowledge could be especially useful for new patients. Patients may also have a negative attitude towards taking medication or experience other barriers towards a correct dosing regimen (Hong et al. 2010; Rees et al. 2010; Stryker et al. 2010; Budenz 2009). Additionally, a recent study showed that patients who visited their ophthalmologist at least every 3 months had a lower risk of nonadherence (Vandenbroeck et al. 2011). Recently, communication techniques have been advocated to identify nonadherent patients and explore their individual barriers to adherence with glaucoma medication (Hahn 2009). These techniques may facilitate interventions to improve adherence, tailored to the individual patient's needs (Budenz 2009). Needless to say, further studies are needed to evaluate the effects of such a more individualized approach. The authors wish to thank all investigators and their staff from the following participating hospitals: T.P. Colen, MD, PhD, Amphia Ziekenhuis; N. G. Tahzib, MD, PhD, S. Gast, MD, E. C. La Heij, MD, PhD, Atrium Ziekenhuis; S. Nobacht, MD, Canisius Wilhelmina Ziekenhuis; M.J.W.M. Busch, MD, PhD, Catharina-Ziekenhuis; F.C.A.P. Dijkman, MD, Deventer Ziekenhuis P.J. Kruit, MD, PhD, Diaconessenhuis Leiden; R.C.W. Wolfs, MD, PhD, Erasmus MC, UMC Rotterdam; A.J.M. van Hogerwou, MD, Gelre Ziekenhuizen; G.W.S. Thoe Schwartzenberg, MD, Groene Hart Ziekenhuis; H.F.A. Duijm, MD, PhD, Isala Klinieken; H.M.A. Brink, MD, PhD, Jeroen Bosch Ziekenhuis; J.P. Kappelhof, MD, PhD, Meander Medisch Centrum; B.S. Hwan, MD, Medisch Centrum Alkmaar; M.A.J. Wagemans, MD, PhD, OMC Haarlem; R.J. Wouters, MD, PhD, Westfries Gasthuis; H.J.B. van den Brom, MD, PhD, Wilhelmina Ziekenhuis, S. van der Veen, MD, Ziekenhuis de Tjongerschans; And all investigators and staff from the Maastricht University Medical Centre. Alcon Nederland B.V., Gorinchem, provided financial support for this study. An abstract of the article was presented at the Dutch Ophthalmological Society meeting, Maastricht, 25–27 March 2010, and at the Vanguard Glaucoma Council, Brussels, 23–25 September 2010.