Money for medication: a randomized controlled study on the effectiveness of financial incentives to improve medication adherence in patients with psychotic disorders

Approximately 60% of patients with psychotic disorders experience difficulties being adherent over time or fail to take their medications as prescribed, with mean non-adherence rates around 50% [1]-[3]. Moreover, among patients who do not openly refuse to accept their antipsychotic medication, many are only partially adherent [4]. Failure to take the medication as prescribed is associated with a wide array of adverse individual and societal outcomes such as inconsistent symptom control, more relapses [5]-[7], more (re)hospitalizations [8],[9], more suicide attempts [10],[11] and more encounters with police and justice, either as a victim or as a perpetrator [12].

Clinical advantages of antipsychotic medication are often limited by patients’ failure to adhere sufficiently to their prescribed medication. This partial compliance severely reduces the effectiveness of the medical treatment of schizophrenia and interferes with therapeutic efforts [13]. For example, relapses occur within one year for 50 to 75% of the patients with schizophrenia after discontinuing with their antipsychotics [14]. Missing antipsychotic medication has also been associated to double the risk for hospitalization [9]. Throughout this protocol alternative definitions such as ‘acceptance’ or ‘compliance’ relate to the concept of medication adherence.

Risk factors for non-adherence have been studied extensively and were systematically reviewed by Higashi and colleagues [15]. They distinguished (1) patient-, (2) treatment-, and (3) environmental-related factors to be associated with non-adherence.

Unfortunately, most studies investigating interventions to improve adherence yield inconsistent results and do not always lead to less symptoms, better functioning or improved quality of life [26],[27]. Therefore, a (combination) of innovative methods is needed to help patients take their antipsychotic medication as prescribed [28]-[30]. One such innovative intervention is contingency management.

Contingency Management (CM) interventions typically reinforce pre-set, well-defined and verifiable target behaviors (e.g., drug abstinence or medication intake), by providing financial incentives or vouchers. Interventions based on CM-principles have been applied in various settings targeting a variety of behaviors, and have shown robust effects in reducing drug use and increasing treatment compliance and medication adherence (for overviews see; [31],[32]). Currently, no studies have investigated the effect of CM for non-depot medication adherence in patients with schizophrenia.

In reviewing studies using CM based interventions in patients with mental health problems, Priebe et al. [33] did not find any randomized controlled studies testing the effectiveness of financial incentives to improve depot medication adherence in patients with psychotic disorders. However, two pilot studies were conducted that showed promising results.

Claassen and colleagues [34] included five non-adherent patients of which four patients accepted financial incentives upon medication acceptance. This resulted in improved adherence rates and significantly decreased patients’ hospital admissions during the intervention period. Staring et al. [35], also included five non-adherent patients with psychotic disorders in their pilot study. Results showed that the percentage of accepted depot injections increased from an average of 44% in the previous year to 100% in the year in which financial incentives were offered. While patients had been hospitalized for an average of 100 days in the preceding year, only one patient was re-admitted for 17 days during the intervention year. More recently, the first cluster randomized controlled trial tested the effectiveness of offering financial incentives to patients (n = 141) with psychotic disorders who were partially non-compliant to improve their medication adherence [36]. Interestingly, although adherence to antipsychotic depot medication increased significantly in the CM group as compared to the control group (85% and 69% acceptance of depot after one year), this did not result in a significant difference on clinician rated clinical improvement. In sum, two pilot studies showed promising results and one RCT showed partial positive results of financial incentives upon acceptance of antipsychotic depot medication.

The goal of the current study is to assess the effectiveness of providing financial incentives upon depot acceptance in psychotic disorder patients. The primary objective of this study is to assess the effectiveness of M4M during the intervention in terms of acceptance of antipsychotic depot medication (the medication possession ratio; MPR). To assess how discontinuing the intervention affects depot acceptance, we will also compare the MPR during the follow-up period (six months), in which no CM takes place. In addition to the MPR, secondary objectives include the longest uninterrupted period of depot medication acceptance, the expired time before depot is taken and attitudes towards medication. Our third objective is to assess the effects of medication acceptance on patients psychosocial functioning, quality of life, cost-utility, substance use and side effects of the antipsychotic medication.

In a parallel-group randomized controlled trial, patients will be randomly assigned to the experimental condition (M4M), or to the treatment as usual (TAU) control condition. Note that during the recruitment phase of the study, only patients who are prescribed or have an indication for antipsychotic depot medication, and who have expressed their willingness to accept antipsychotic depot medication are eligible for inclusion and after providing written informed consent for randomization. Patients assigned to the experimental condition (M4M, n =84) will receive a financial incentive for each prescribed depot they accept, in addition to treatment as usual. Patients in the control condition (TAU, n =84) will receive treatment as usual only without financial incentives upon depot acceptance. After randomization, both patients in the TAU condition and patients in the M4M condition are prescribed depot medication. After the intervention period of 12 months, there will be a follow-up period of 6 months in which patients in both study groups receive TAU and no financial incentives for accepting their prescribed antipsychotic depot medication.

Patients will be 168 outpatients with a psychotic disorder from three mental health care institutions in the Netherlands: (1) Palier (‘Dual Diagnosis Centre’ (CDP)), (2) Parnassia and (3) BavoEuropoort. These organizations primarily treat patients with psychotic and other severe mental disorders, (often with comorbid substance use disorder), from the cities of Rotterdam and the Haque in the Netherlands. Per team around two hundred patients with a psychotic disorder are treated. Patients will be recruited based on the following inclusion criteria: age between 18 – 65 years, a psychotic disorder (including schizophrenia, schizoaffective disorder or other psychotic disorders), taking antipsychotic depot medication or an indication to start using depot medication, outpatient treatment (either starting outpatient treatment after discharge from a psychiatric hospital, or being in outpatient treatment for at least four months), and given informed consent. In concordance with their psychiatrist, patients who will start using antipsychotic depot medication are considered to have an indication for antipsychotic depot medication. These patients are - if they meet the other inclusion criteria- eligible to contact for our study. Exclusion criteria are the inability to participate due to cognitive impairments and/or insufficient understanding of the Dutch language (clinical judgment). Refer to Figure 1 Participant flowchart for details.

Patients assigned to the intervention group (M4M; Money for Medication) will receive treatment as usual (see below), plus a financial incentive for each time they accept their prescribed depot of antipsychotic medication during the 12 months experimental study phase. All patients in the M4M group will receive a maximum of 30 euro per month. The amount of money per accepted depot is dependent upon the frequency of depot administration. For example, a patient who receives one depot every two weeks will receive 15 euro per accepted depot. A patient who receives one depot every three weeks will receive 22.50 euro for each accepted depot, et cetera. Patients receiving oral penfluridol with a frequency of once a week will also be included in the study. Oral penfluridol is used when patients have problems taking antipsychotic medication on a daily basis, and when they do not accept intramuscular depot injections. They will receive 7.50 euro for each time they accept their weekly penfluridol oral tablet. The financial incentives will be given by the patients’ treating nurses directly upon administration of the depot or penfluridol. They will receive their depot medication primarily at the clinic (‘depot room’) and sometimes at home during home visits. Patients will sign a proof of receipt.

Patients assigned to the control group will receive treatment as usual (TAU) during the 12 months experimental study phase and during the 6 months follow-up. TAU includes outpatient treatment provided by community mental health teams and flexible assertive community treatment teams [37]. All clinicians encourage continuing depot medication in case this is prescribed by the psychiatrist of the team. When needed, crisis services can be used or patients can be hospitalized (in)voluntarily. The type and dosage of the depot antipsychotic medication and other medications patients receive will be determined by the patients’ psychiatrist together with the patient. The type, frequency and dosage will not be affected by participation in the study. Administration of the depots will be done by the psychiatric nurses working in the teams.

Candidate participants will be selected from the caseloads, applying the in- and exclusion criteria. Patients who meet the criteria will be informed and asked to participate by their clinician. Patients who consider participation receive a take-home brochure with information about the study. The clinician asks the patient’s permission to be contacted by a researcher. If the patient agrees, the researcher contacts the patient to schedule an appointment for the baseline interview. If a patient indicates that he or she does not want to participate, this will be registered anonymously together with their demographic and clinical characteristics (DSM IV-TR diagnosis on axes I and II) to enable assessment of selection bias. If possible, the patient will be asked to explain why he or she does not want to participate.

With support of the management, all teams and their clinicians have expressed their willingness to co-operate with the conduct of our Money for Medication study. Clinicians of course can be resilient about the concept and intervention of our study. Therefore, we assess clinicans’ attitudes towards M4M.

Prior to the baseline interview the researcher explains the design and purpose of the study, the research goals and the randomization procedure. After written informed consent is given, the baseline interview will take place and subsequently, participants will be randomized to the intervention (M4M) or control condition (TAU). Randomization will be stratified by site, gender, substance use disorder (absent vs. prevalent) and previous compliance with antipsychotic medication (compliance rate <50% vs. ≥50%). There will be three interviews at 0, 12 and 18 months (see Table 1). All participants will receive a remuneration of 20 euro for each interview. In the cases where the researchers cannot overcome certain practical obstacles (e.g. imprisonment, hospitalization), patients who can demonstrate that they have accepted their depot medication (for instance in the form a written statement by the treating prison or hospital medical doctor), receive their monetary reward as soon as possible, but with a delay. In case of discontinuation of depot intake, the monetary reward will stop and data of non-depot medication intake will be monitored in order to have a complete overview on the number of patients discontinuing depot medication and switching to non-depot.

Originally the start date for patient recruitment was May 21, 2010 and was planned to be completed by September 2012. Due to a change in personnel and organizational factors that caused logistical delays, patient recruitment was low and continued again in September 2013. Note that this is an ongoing study and that we expect to finish our complete data collection by April 2016. Therefore, we expect to submit the results of this study in 2016.

The study protocol has been approved by the accredited Dutch Medical Ethical Trial Committee (METC) of the Erasmus University Medical Centre (registered under number NL31406.097.10 and file number P13.258). According to the Dutch Data Protection Act (DPA) data will be safely stored and anonymized and is only accessible for members of the research group or the Medical Ethical Committee. All patients will provide informed consent before entering the study.

Following the CONSORT statement we calculated our power to the primary outcome measure of this study and not for the secondary or tertiary outcome measures. Based on previous findings and study protocols [33],[60], we expect a difference of 65 to 85 percent (an absolute difference of 20%) of accepted depots between the control group and the money-for-medication condition. In terms of Cohen’s h, this constitutes a medium effect size (h =0.5). With Type I error rate (alpha) set at 5%, power at 0.90 (1 – Type II Error rate; 1-β), and 20—25% drop-out, we will need 84 patients per arm to detect an absolute difference of 20% [61]. In total 168 patients will be included.

The primary outcome will be reported as accepted depots as percentage of planned depots, most often weekly, biweekly or monthly. The effects of the intervention on our outcome measures will be analysed using generalized linear models as appropriate to the outcome, with random effects for sites or treatment teams. Sensitivity analyses will be conducted to explore the impact of different strategies for handling missing data. A detailed analysis plan will be completed prior to analysis of baseline measurements.

In the pilot study of Claassen et al. [34], M4M did not have a negative impact on the therapeutic relationship. Furthermore, they have not found that other patients who did not participate in the M4M study complained about unequal treatment or demanded to be paid for taking medication as well.

In the pilot study of Staring, Mulder, and Priebe [35], all five participants considered M4M to be a good project. The reasons they gave to participate in the pilot were “I don’t like the injection, but money makes it better”, “Money keeps me motivated”, and “The depot injections keep me balanced”. When prompted, two patients said that they perceived financial incentives as a voluntary and non-coercive measure, two patients did not know what to think about this, and one indicated that he perceived financial incentives as a coercive measure, saying that “I have to take the medication anyway”. All patients said that they spent the money on food and cigarettes, and one patient also bought household products. It was observed, however, that at least one patient had spent some of the money on cannabis. Other patients did not ask to be offered incentives as well and no negative impacts on therapeutic relationships were noted. Some patients however felt that they should receive more money (they received 10 euro for every two weekly depot, 15 euro for every three-weekly depot and 20 euro for every four-weekly depot).

Although higher payments have been found to result in bigger effects [62],[63], in the present study we have chosen a maximum of 30 euro’s on average per month because (1) patients could get used to or become financially dependent on higher payments or even lose their social security benefits and (2) to strive for acceptable cost-utility of M4M. From a societal perspective it is important to focus on cost-effective interventions and given the results of prior research [35], it seems not necessary to use higher incentives.

Our study started in 2010, prior to some of the recent findings as described above. The difference with the earlier RCT studying the effects of M4M [36] is that we will include both patients who are partially non-compliant, as well as patients who are compliant in taking depot medication. The rationale to also include compliant patients is the observation in several studies that around thirty percent of patients initiated on antipsychotic depot medication cease to accept their depot within one year [5],[64],[65]. In addition, when we eventually might want to implement this intervention into daily clinical practice, it is more ethical as well as more practical to reward both compliant and non-compliant patients.

A possible limitation of this study are the different medication depots. Although most patients receive medication injections it is also allowed for patients to take oral penfluridol. The disadvantage of this oral antipsychotic medication is that it is more difficult to check if patients actually take their medication. However, penfluridol is also a depot and because we aim to test if our intervention is broadly applicable we decided to include both patients with oral penfluridol and injections.

Another limitation is that the clinicians cannot be blinded to the intervention condition, possibly resulting in a more stimulating attitude for accepting depot medication in the intervention group. Also, the interviewers are not blind to the patients’ condition and therefore can rate patients’ responses as more positive or negative.

Furthermore, the intervention effect can be influenced by different depot frequencies. For example, patients who receive money every week are rewarded four times as often compared to patients that receive depot every month. Receiving a small incentive more frequently can be more stimulating or motivating compared to receiving a bigger incentive only once a month. This can interfere with our intervention effect, even though the mean amount of money per month remains equal for all participants.

Ethical concerns have been raised about paying patients to accept their medication and whether this is an acceptable means in the treatment of patients with psychotic disorders [66],[67]. One of these concerns is that patients’ intrinsic motivation to accept medication will disappear if money is involved. We will study this by assessing intrinsic motivation over time, as possible decreases in depot acceptance can occur during the 6 month follow-up without M4M. Another frequently raised ethical argument is that patients might buy drugs or alcohol from the money they receive. We will monitor alcohol and drug use by using assessment scales as well as obtaining urine samples.

Apart from these ethical concerns, we will also assess the intervention from a cost-utility perspective, because this is an important factor to consider from a societal point of view. In conclusion, we will test if M4M improves patients’ MPR, reduces their psychotic symptoms and contributes to a clinical improvement.