Effect of an in-hospital medication self-management intervention (SelfMED) on medication adherence in polypharmacy patients postdischarge: protocol of a pre-post intervention study


The WHO states that 50% of patients afflicted by chronic conditions do not adhere to prescribed medication regimens.1 After hospital discharge, 24%–70% of patients are found to be non-adherent to their medication.2–5 Furthermore, polypharmacy is found to be associated with poorer medication adherence due to the larger number of medicines and the complexity of the medication regimen.5–8 Such medication non-adherence can result in poor health outcomes (eg, increased mortality, decreased quality of life and loss of productivity), increasing healthcare service utilisation and healthcare expenditures.9–15

Self-management of medication (SMM) is one of the promising strategies to improve patients’ medication adherence and self-care competencies, as well as a method to increase patient satisfaction and patient safety.16 17 SMM comprises a range of tasks that individuals must successfully perform to self-manage their prescribed medication regimen and entails knowledge, skills and behaviours mandatory to correctly take medication and sustain use over time.18 According to Bailey’s model, the medication self-management process consists of six steps: (1) filling prescriptions and picking up medicines at the pharmacy; (2) knowledge and understanding of medicines; (3) practical organisation and scheduling of medicine intake; (4) medication intake; (5) monitoring therapeutic and adverse effects and (6) sustaining safe and appropriate medication use over time.19 Failing the process of medication self-management can affect medication adherence. For instance, forgetting to take medicines, being unable to follow medication-related instructions or being unable to open medication packages may result in non-adherence.20 Therefore, interventions tailored to increase medication self-management are expected to improve medication adherence.21

The idea of in-hospital SMM consists of patients who store, prepare and administer their medication themselves. During the process of in-hospital SMM, nurses, physicians and hospital pharmacists monitor and support patients, provide education and counsel on supportive measurements. If problems arise, the multidisciplinary team can identify and address issues and prevent medication-related problems after discharge. SMM during hospitalisation may provide patients an opportunity to continue medication management routines from home.

However, the number of patients allowed to self-manage their medication during hospitalisation is currently limited as hospital procedures are designed for medication management by healthcare providers. A multicentre study performed in Belgium demonstrated that although up to 41% of patients would be able to self-manage their medication during hospitalisation, only 22% did self-manage one or more medication(s).22 This prior study demonstrated only 10 out of 57 wards had a procedure for SMM and 4 out of 57 wards had a screening tool to assess the competences of patients to self-administer their medicines. When SMM was allowed and supported during hospitalisation, it seldom occurred in a structured, evidence-based way.22

Hence, an evidence-based intervention (ie, medication self-management intervention, SelfMED) supporting in-hospital SMM was developed and tested.23 The SelfMED intervention consists of four components: (1) a procedure clarifying the SMM process and affiliated responsibilities (ie, SelfMED procedure); (2) an assessment evaluating patients’ eligibility for SMM during hospitalisation (ie, SelfMED assessment); (3) a monitoring system allowing healthcare providers to follow adherence and detect errors in patients’ SMM (ie, SelfMED monitor) and (4) measures for healthcare providers to support SMM in patients (ie, SelfMED support). A paper-and-pencil version of the SelfMED intervention has been pilot tested at one cardiology unit of a general hospital. Pilot-based results indicated the intervention facilitates in-hospital SMM and was found to be safe. However, an important knowledge gap exists on the effectiveness of the SelfMED intervention on patients’ outcomes after discharge.

Literature on the effects of in-hospital SMM suggests an increase in patient adherence, medication knowledge and patient satisfaction.16 17 24–26 These findings are, however, based on a limited number of studies and the overall quality of studies was rather low making it hard to draw any firm conclusions. In addition, these studies focused solely on the short-term effects of SMM (eg, during hospitalisation, at hospital discharge, 2 weeks postdischarge). To date, no study has been conducted investigating the effects of SMM on various outcomes 2 months postdischarge in patients with polypharmacy.

The primary objective of this study is to evaluate the effect of in-hospital medication self-management (ie, SelfMED intervention) on medication adherence in patients with polypharmacy 2 months postdischarge, as compared with usual care (ie, medication management by healthcare providers). Secondary objectives are to evaluate the effect of in-hospital medication self-management on outcomes at the patient level (ie, self-management, knowledge, satisfaction), at the healthcare provider level (ie, staff satisfaction, workload) and at healthcare system level (ie, healthcare services utilisation).

The primary hypothesis is that, compared with usual care, the SelfMED intervention improves medication adherence at 2 months postdischarge. The secondary hypotheses are that patients’ medication self-management, medication knowledge and patient and staff satisfaction improve while workload and healthcare service utilisation will decrease.


Study design

A multicentre prospective pre–post intervention study will be conducted from February 2024 to February 2025. The study will start with a 4-month control phase investigating usual care. Medication management of patients included during this phase will be entirely provided by healthcare providers. The control phase is followed by a 4-week period to train and prepare healthcare providers on the SelfMED intervention. Subsequently, the study will proceed with a 4-month intervention period investigating the effects of the SelfMED intervention with rigorous monitoring and support of medication-taking behaviour during hospitalisation. Patients recruited in both the control and intervention periods will be monitored over a course of 2 months postdischarge. Figure 1 presents a schematic illustration of the study design.

Figure 1
Figure 1

Schematic illustration of study design.

A pre–post test design was selected as random allocation of patients to the study groups at one ward is associated with a high risk of contamination bias (ie, anticipated effect of the intervention on usual care provided to the control group). Furthermore, this design reduced the risk of safety issues as healthcare providers would have to follow different procedures for different patients. Moreover, a stratification in intervention and control wards has also been rejected due to potential ward-related bias.

Study setting

Patients will be recruited in two Belgian hospitals (ie, Antwerp University Hospital and Jessa Hospital). Hospital wards will be chosen based on the presence of patients with chronic conditions and associated polypharmacy (eg, oncology, cardiology and diabetology).

Study population

Dutch-speaking patients (≥18 years), taking ≥5 chronic medications at the time of hospital admission (ie, polypharmacy) of which at least three are administered orally, expected to be hospitalised for at least 3 days and a planned discharge towards home, will be eligible. A medication is classified as chronic when it is intended to be taken consistently for a minimum of 3 months (minimum expected treatment period). Nursing home residents, patients with a neuropsychiatric diagnosis of dementia or Mini Mental State Examination <24/30, patients with a terminal illness (ie, life expectancy less than 3 months), and those unable to provide consent, will be excluded. An additional criterion for patients being recruited consists of being eligible for medication self-management according to the SelfMED-assessment tool (see further).23 All patients meeting the inclusion criteria and willing to participate will be included (consecutive sampling).

Sample size

Sample size estimation is based on the primary outcome, more specifically medication adherence at 2 months postdischarge measured using pill counts. The intervention is deemed successful if the medication adherence rate is significantly higher in the intervention group than in the control group. A systematic review and meta-analysis on medication adherence outcomes of 771 intervention trials showed a standardised mean difference of 0.290 comparing treatment and control groups27 and is, therefore, considered as effect size for power calculation in this study. This rather small effect size was used, so that even small differences in adherence can be detected, although a larger effect of the intervention is expected.

Power calculation using G-power showed that, to detect an effect size of at least 0.29, with an alpha of 0.05 and a power of 0.80, a minimum sample of 374 participants would be needed (n=188 per group). Previous research on the evaluation of measurement methods for medication adherence in patients with polypharmacy for 2 months showed a drop-out rate of 19%.28 Hence, to provide a sufficient margin, we assume 20% attrition. In summary, the recruitment of 225 patients in each group will be needed to achieve the desired sample size.

Control period (preintervention)

During the control period, care as usual will be provided comprising medication management (ie, storage, preparation and administration) entirely provided by healthcare providers during hospitalisation. Patients are not actively involved in the medication process and do not administer their medication themselves during hospitalisation.

On admission, a medication anamnesis is conducted to complete and verify the patient’s medication list (home medication). During hospitalisation, physicians electronically prescribe medication, considering the medication list and the hospital’s formulary. All medications for in-hospital are dispensed by the hospital pharmacy. Most hospitals use a mixed drug distribution system, with ward stocks for common medications and patient-level dispensing from the central pharmacy. Nurses are responsible for medication administration and documentation of the administration in the electronic medical record. On discharge, patients are provided with an updated medication list along with information about any newly initiated or modified pharmacotherapy. If necessary, patients receive prescriptions for newly initiated or modified pharmacotherapy, which should filled by the outpatient pharmacy.

Healthcare providers’ training

Healthcare providers involved in the study will receive a 1.5-hour training session from the first author (LM). During this session, healthcare providers will be informed of the study objectives and detailed intervention protocol. Healthcare providers will learn how to complete the SelfMED assessment to ensure uniformity. They will gain insight into the prerequisites and practical organisation of medication self-management by patients during hospitalisation. Furthermore, it will be explained how to use the SelfMED monitoring system and SelfMED support tool (supportive measures).

To give everyone the opportunity to attend a training session, considering work schedules, holidays etc, multiple training sessions will be organised over a 4-week period. Healthcare providers can attend sessions multiple times if needed. At least all full-time working nurses should attend the training to ensure the correct implementation of the intervention. Healthcare providers will also receive a written standardised operation procedure. The PhD researcher (LM) or a trained research assistant will visit the hospital ward(s) daily to support healthcare providers in implementing the intervention components, including performing the SelfMED assessment, monitoring SMM and supporting SMM.


During the intervention period, SMM during hospitalisation (ie, SelfMED) will be provided with rigorous control of medication-taking behaviour, including education and support from healthcare providers to improve pharmaceutical self-care. SelfMED is an evidence-based intervention consisting of multiple components23 (see figure 2).

Figure 2
Figure 2

Overview of subsequent interventional SelfMED components. SelfMED, medication self-management intervention.

(1) SelfMED assessment: A validated stepped assessment tool evaluating patients’ eligibility for SMM during hospitalisation.23

  • Nurse assessment: A 10-statement assessment allows the nurse to assess patients’ eligibility for SMM on admission. Completing this assessment is performed based on information gathered during the intake process and data accessible in the patient’s electronic medical records. Nurses can indicate their response to the statement in the first column (ie, agree/not agree/not known) and evaluate whether the statements pose a barrier to self-management in the second column (ie, barrier/possible barrier/no barrier to self-management). At the end of the assessment, the nurse indicates if the patient is eligible for self-management and subsequently invites the patient to complete the patient self-assessment.

  • Patient self-assessment: If the nurse deems the patient capable of self-managing medicines, the next phase of the procedure consists of a patient questionnaire. The questions in the assessment consist of the current medication management at home (one item), the patient’s willingness to self-manage medication in hospital (one item), a possible need for aide while self-managing in hospital (one item) and the patient’s therapy adherence out of hospital (seven items). If patients indicate they are not self-managing medication at home or are not willing to self-manage medication, they do not have to complete the questionnaire.

  • Nurses’ advice: The information from both the nurse’s assessment and the self-assessment enables nurses to advise the treating physician to allow or decline medication self-management. If the nurse does not recommend SMM, there is a possibility to advise a reassessment at another moment in time since improvement in health status can result in improvements in self-management capability.

  • Physician assessment: In the next phase, the patient’s physician has to make the final decision on allowing the patient to self-manage medication. The advice formulated by the nurse, the nurse’s assessment and the patient’s self-assessment are available to make an informed decision. If the treating physician decides to not allow medication self-management, there is a possibility to reassess the patient at another time point. If a patient is deemed eligible and self-management is allowed, the actual self-managed medicines need to be specified by the treating physician and registered in the patient’s electronic medical file. The physician can, for example, decide not to allow SMM for specific acute or high-risk medications. Intravenous medication or medication that patients typically do not administer themselves in the outpatient setting will be excluded from SMM and administered by nurses during the standard medication rounds.

(2) Starting SMM: If a patient is deemed eligible for SMM, several practical arrangements must be made to enable the execution of SMM:

  • The self-managed medication needs to be documented in the patient’s electronic medical file.

  • The patient should be provided with an up-to-date personalised medication list.

  • The hospital will provide all inpatient medication according to the hospital’s formulary. The medication should be delivered by the hospital pharmacy to the patient’s room. Therefore, X number of medicines should be ordered from the hospital pharmacy, so that the patient has a possible supply of medication, simulating the home situation. The practical organisation of medication delivery to the patients’ room will be determined and aligned with common procedures in each hospital in consultation with the hospital pharmacist(s) prior to the start of the study (eg, for how many days medication will be delivered as well as the mode of delivery).

  • The patient needs to be educated on the self-managed medication and the use of the medication list. Since the hospital will provide all inpatient medication according to the hospital’s formulary, medication will be substituted to a different brand or pharmacotherapeutic class (if a home medication is not on the hospital’s formulary). Therefore, extra attention should be paid to educating patients about switching medication.

  • The patient needs to be educated on the use of the SelfMED-monitoring system (see further).

  • Medication should be safely stored in the patient room (preferably in a lockable locker, personal closet or beside table). Possibilities of medication storage will be discussed in advance with the patient care department of each hospital.

(3) SelfMED monitor: A monitoring system allowing healthcare providers to follow up adherence and detect errors in patients’ SMM. As healthcare providers monitor patient adherence and safety, patients have to register per medicine the dose and timing of the medication intake on a daily base. For medication registration, a paper-based medication list is used similar to the one provided to the patient at discharge. It allowed the patient to use additional aids such as a clock or an alarm to remind him of the medication intake. Nurses are required to check the registration form twice a day, that is, during both the early and late shifts, to identify any discrepancies with the prescribed regimen. If problems concerning medication self-management occur or there is doubt about medication intake during these two control moments, a pill count can be performed optionally to detect medication errors. Discrepancies and self-management problems experienced will be discussed with the patient.

Since medication management should be monitored twice a day, the monitoring file should also be completed twice. The monitoring tool consists of a first column to define the date and initials of the monitoring nurses that day. The second and third column questions whether the patient is still capable to self-manage medication. If not the reason should be formulated (eg, sudden illness and mental decline). If the status of the patient has not changed and he/she was self-managing medication, the medication list is then evaluated, and a pill count is optionally performed. The results should be noted in the monitoring file in the patient’s room. If the patient succeeds in every aspect of self-management, self-management can be continued. If patients fail to self-managing their medication, the nurse can intervene by providing patient education to prevent this error from occurring again. The intensity and duration of the action will depend on the specific problem encountered. If the errors in the self-management are found to be problematic, it is possible to stop self-management. In case a patient is not capable of self-managing medication anymore, self-management stops. A possibility to reassess over a period of time is then provided.

(4) SelfMED support: A guidance document for healthcare providers to support patients with SMM was developed based on scientific evidence and expert consensus. The guidance document consists of a list of possible problems concerning medication self-management and specific recommendations to address these problems. The guide can be used as a resource for healthcare providers to take action in consultation with the patient.29 To fully support patients in medication self-management, the advice from the guide will also be made directly available to patients through an information leaflet.

Randomisation and blinding

Given the nature of the intervention, blinding is not possible. Since usual care comprises medication management provided by healthcare providers, patients in the intervention group will notice a change in the method of medication administration. Therefore, it is not possible to blind participants or healthcare providers. There will be no random allocation of patients to an intervention or control group (cfr. study design).

Outcome measures

Figure 3 provides an overview of the outcome measures and measurement time points. All outcomes will be evaluated in participants of both the control and intervention groups, except for healthcare providers’ satisfaction and workload.

Figure 3
Figure 3

Overview of data collection and outcomes. HCP, healthcare professional; SelfMED, medication self-management intervention.

Primary outcome

The primary outcome consists of ‘medication adherence’ measured by pill counts, based on preserved medication packages, and determined at 2 months postdischarge. A threshold of 80% of medications taken will be used to differentiate between adherent and non-adherent patients.30

Secondary outcomes

  • Patient-reported medication adherence will be measured in addition to pill counts, as the best strategy for measuring adherence appears to be the combined use of at least two methods.30 31 A self-report questionnaire (ie, Probabilistic Medication Adherence Scale—ProMAS) will be completed by patients on admission and 2 months postdischarge.32 During 2 months, patients are requested to keep a structured medication diary in which medication-related problems are registered (on paper).33 Problems could be, for example, forgetting to take medication or not taking medication because of side effects. The feasibility and validity of the combination of pill counts, a medication diary and a self-report questionnaire have been pilot-tested28

  • SMM will be defined as ‘the patients’ ability to manage their medication regimen’.18 On admission and 2 months postdischarge, problems with medication self-management experienced by patients will be evaluated through a self-report survey (see online supplemental file 1). Problems with medication self-management will be surveyed using statements to be answered on a 5-point Likert scale (1=always, 5=never). The medication self-management problems in the self-developed survey are divided according to the six phases of Bailey’s medication self-management model19 and are based on a list of medication self-management problems compiled in previous research.29

  • Medication knowledge: On admission and 2 months postdischarge, four types of medication will be randomly selected from the patient’s medication list. Regarding these medications, knowledge will be assessed including medication name, indication, dose, administration route and administration time (correct yes/no). The sum scores (=sum of all correct answers) of total knowledge of chronic medication will be calculated.

  • Patient satisfaction with pharmaceutical care: At discharge, patients will be surveyed about their satisfaction with the received pharmaceutical care during hospitalisation through a Visual Analogue Scale.34

  • Healthcare provider’s satisfaction with pharmaceutical care: At the end of the study, nurses, physicians and pharmacists will be surveyed about their satisfaction with pharmaceutical care provided through a visual analogue scale, as well as the advantages and disadvantages of usual care and medication self-management using open-ended questions. Free-text comments will allow healthcare providers to list the (dis)advantages of both working methods.

  • Healthcare providers’ workload: At the end of the study, nurses, physicians and hospital pharmacists indicate the impact of pharmaceutical care on their workload using a 10-point rating scale.

Supplemental material

  • Healthcare service utilisation: Two months postdischarge, patients will be questioned about medication-related advice-seeking behaviour and health services utilisation through a self-reported questionnaire including number of hospital readmissions, nurse visits and consultations with general practitioners.

Potential covariates

At baseline, participants’ sociodemographic characteristics will be collected. Information about age, gender, highest educational level, type of hospitalisation (acute/planned), reason for hospitalisation and the number and type of prescribed medications will be collected. To gather information on medication, we will copy the medication list on admission and at discharge.

Compliance with the different interventional components will be evaluated per participant at the end of the study based on available data from the documents completed by healthcare providers. This includes determining whether the SelfMED assessment was completed by the nurse, patient and physician; checking if daily monitoring was performed according to the SelfMED monitor/follow-up form, and verifying whether action was taken if problems with medication self-management were identified according to the follow-up form.

Data collection

Data collection will proceed in several steps. First, patients’ eligibility will be evaluated by the (head)nurse of the participating wards based on the inclusion and exclusion criteria. If patients are eligible, the nurse assessment will be performed. If nurses deem the patient eligible for medication self-management, patients will be invited to participate in the study. If patients refuse to participate, reasons for non-response will be surveyed. Patients willing to participate will be asked to sign informed consent, and subsequently to complete the patients’ SelfMED self-assessment. This assessment will be extended with a questionnaire on demographics (ie, age, gender, educational level, type and reason for hospitalisation) and medication intake at home. Afterwards, the treating physician will complete the physician assessment. Further participation in the study is only possible if the physician decides that the patient is eligible for in-hospital medication self-management. The SelfMED assessment will be performed in both the intervention and control groups.

On admission (after inclusion), a baseline measurement will be performed in which data on medication adherence, medication self-management and medication knowledge will be collected in both the intervention and control groups (see the ‘Outcome measures’ section). During hospitalisation, in the intervention group, the SelfMED monitoring tool will collect data from the patients’ log of the medication scheme and the nurse’s pill count. At discharge, the medication schedule of patients from both groups will be copied to collect data about medication use. Furthermore, data on patient satisfaction with pharmaceutical care will be collected in both groups (see the ‘Outcome measures’ section). At discharge, patients from both the intervention and control groups receive oral and written instructions on the use of the medication diary and the preservation of their medication packages for a 2-month period (±60 days). Two months postdischarge, a home visit will be conducted in both the intervention and control groups by the first author or a trained research assistant. During this home visit, several questionnaires will be administered to collect data on medication adherence, medication self-management, medication knowledge and healthcare service utilisation (see the ‘Outcome measures’ section); the preserved medication packages will be collected to perform a pill count and the medication diary will be collected for analysis. Approximately 2 weeks before the home visit, participants will be contacted by phone to remind them of the upcoming appointment. At the end of the study, healthcare providers will be questioned about their pharmacology care workload and their satisfaction with pharmaceutical care (both usual care and medication self-management).

To ensure uniformity of data collection, the administration of questionnaires will be conducted by the first author or trained research assistants. Through an information session, research assistants will be informed about the instructions (medication diary and medication packages) to be given to patients and the administration of questionnaires on admission, at discharge and 2 months postdischarge. Data will be collected by means of paper questionnaires.

Statistical analysis

Data will be analysed by using IBM SPSS Statistics V.29. Descriptive analysis of the study population will be performed with a comparison of the characteristics of the preintervention and postintervention sample. Discontinuous data will be described using frequency distributions. Continuous data will be described using a mean value and SD if normally distributed or using a median and range if non-normally distributed. Z-scores and Shapiro-Wilk test will be used for testing the normality of the data. χ2 test, Fisher’s exact test, independent t-test or Mann-Whitney U tests will be used to measure differences. A p≤0.05 will be considered as statistical significance, all tests will be performed two-sided.

Medication adherence measured by pill counts, the main outcome, will be measured using the following formula: Embedded Image

. Both the average overall adherence rate and the adherence rate per medicine will be calculated. A threshold of 80% of medications taken will be used to differentiate between adherent and non-adherent patients. Differences in medication adherence between the preintervention and postintervention group will be tested using independent t-tests or Mann-Whitney U tests (in case of adherence rates as dependent variable) or using Fisher’s exact test (in case of the categorical variable adherent/non-adherent as the dependent variable).

To adjust for the possible imbalance between the preintervention and postintervention sample, a multivariable linear regression analysis will be performed, with medication adherence 2 months postdischarge as dependent variable and the group allocation and potential confounders as independent variables. Possible confounders based on differences in baseline characteristics between patients included in the preintervention and postintervention periods will be entered consecutively into the model. 95% CIs will be displayed when reporting the linear regression results.

Secondary outcomes (ie, patient-reported medication adherence, medication knowledge, self-management, nurses’ and patients’ satisfaction, perceived workload, health service utilisation) will be continuous outcomes (ie, the number of problems reported in the medication diary; the sum score on the ProMAS questionnaire, the sum score of total medication knowledge; the number of medication self-management problems; satisfaction and workload ratings on a 10-point scale, the number of readmissions/ consultations/nurse visits). Differences in secondary outcomes between the preintervention and postintervention group will be tested using Independent t-tests or Mann-Whitney U tests.

Patient and public involvement

In preparation for the development of the SelfMED intervention, patients were surveyed about their willingness towards in-hospital medication self-management and its prerequisites.35 36 During the development of the SelfMED support tool (ie, guidance with recommendations for healthcare providers), a panel of patients was involved and asked to assess the usefulness of the recommendations.29 The feasibility of measurement methods for measuring the primary outcome (ie, medication adherence) was explored in a sample of 69 patients, which informed our choice of measurement method.28

Ethics and dissemination

This study will be conducted in accordance with the Declaration of Helsinki.37 All participants will sign written informed consent prior to participation, obtained by the PhD researcher (LM) or a trained research assistant. Data will be pseudonymised and the keys to coding will be stored separately in an encrypted file. The research protocol was approved by the ethics committee of the Antwerp University Hospital (reference no: B3002023000176). Any modifications to the protocol, which may impact the conduct of the study, may affect patient safety, including changes in study objectives, study design, patient population, sample sizes, study procedures or significant administrative aspects will require a formal amendment to the protocol. Such amendment will be approved by the Ethics Committee of Antwerp University Hospital prior to implementation.

Scientific valorisation will be achieved through peer-reviewed publications and conference presentations. Authorship will be based on the criteria of the International Committee of Medical Journal Editors (ICMJE). Summaries or reports in layman’s terms will be disseminated to healthcare professionals through their professional organisations. Deidentified datasets will be available from the authors on request.

This post was originally published on https://bmjopen.bmj.com