What is known about the topic?

Given the shortage of nursing staff, home care must be efficiently organized to continue providing care to the increasing number and proportion of older people in the population. Using AHMD increases medication adherence, and patients accept the dispenser as reliable, easy to use, and helpful in coordinating personal medication management. With properly implemented AHMD, fewer home care visits are needed, which helps solve efficiency challenges.

What does this article add?

The Tailored Implementation for Chronic Diseases checklist, supplemented by the implementation factors of Grol and Wensing, was a useful framework to gain a comprehensive overview of the determinants (which were labelled as barrier, facilitator, or both) for successful implementation of AHMD. A list of practical strategies for implementation was developed, which can be consulted to develop adequate tailored strategies for implementing AHMD in home care. Although the most frequently mentioned facilitators focus on the nurses’ behavior (assessing eligibility, motivating patients, giving instructions, and having sufficient knowledge and confidence), the most frequently mentioned barrier was the patient's behavior (unplanned visits when patients do not withdraw the medication in time).

INTRODUCTION

Older people with chronic diseases require drug therapy to reduce symptoms. Therapy adherence is essential for successful medical treatment.1,2 Non-adherence to drug therapy is significantly associated with all-cause hospitalization and mortality in older people.3 Nevertheless, older people are susceptible to non-adherence due to drug-related factors, such as dosing regimen, side effects and polypharmacy, and patient-related factors, such as cognitive function, health literacy, multimorbidity, and functional decline.3–6 Nonadherence is determined by interactions of medical, personal, and economic factors, the relationship with the physician, and cognitive status.7,8

Older people live longer in their own homes9,10 and health care services are being shifted from hospital care toward home care.11,12 Moreover, nursing shortages are increasing globally.11,13 For these reasons, the Dutch government encourages the use of technologies that save time.14–16

Home care nurses support patients’ medication adherence and prevent errors by providing reminders or handing out medication,17 sometimes multiple times a day.18 Automated home medication dispensers (AHMD) are available as e-health devices to support patients who are unintentionally non-adherent.19,20 The device involved in this study is the Medido AHMD (Vitavanti Healthcare Solutions, Rijswijk, The Netherlands). The medication is provided by a pharmacy in a unit-dose system, consisting of sachets with the prescribed doses of medication for a specific patient at a specific time point.21 Home care nurses place the sachets in the device every 1 or 2 weeks, instead of visiting the patient at every scheduled time point. At a scheduled time point, the AHMD provides an audio signal to remind the patient to take the medication. By pressing a button on the device, the device delivers the sachet(s) containing the prescribed medication for that time point. The device makes a small incision in the sachets so that patients can easily open them. If the button is not pressed within an adjustable scheduled time limit, home care is notified (alarmed). Many settings on the device, such as the sound volume and the length of the incision, are adjustable, for which the Medido helpdesk can be contacted by home care nurses.

A study found that the mean medication adherence rate significantly increased from 49% to 97% after 6 months of using AHMD.19 Furthermore, patients found AHMD reliable, easy to use, and helpful.22 In addition, fewer visits were required.23 A reported disadvantage of AHMD is that the focus on empowerment and self-care is not suitable for all patients.23 The persisting challenge for e-health is that traditional care remains essential for many patients: human hands are perceived as warm and compassionate; technology comes across as cold and impersonal.23 Studies commissioned by the Dutch government15,16 show that the use of AHMD saves home care nurses’ travel time and caring time. Nevertheless, unplanned visits may be necessary due to alarms or technical malfunctioning of an AHMD (unknown frequency). According to these studies, approximately 5500 AHMD were in use in the Netherlands in 2021, out of a potential of 28 500.

To determine adequate strategies to implement an e-health innovation such as AHMD, barriers and facilitators in home care must be considered to tailor the implementation strategy to this particular context.24 The most frequently mentioned barriers for implementing e-health in general are limited exposure/knowledge of e-health, lack of necessary devices, and problems with financing.25 The most frequently mentioned facilitators are ease of use, improvement of communication, and motivation.25 However, the barriers and facilitators for implementing AHMD specifically have not yet been investigated.

AIM

The current study aimed to explore the barriers and facilitators for implementing AHMD in home care from the perspective of home care nurses in the Netherlands. Home care organizations can use this knowledge to develop tailored strategies for implementing AHMD.

METHODS Design

A descriptive qualitative study was conducted, using semi-structured interviews. The Consolidated Criteria for Reporting Qualitative Research (COREQ)26 checklist was followed to ensure transparent reporting (Appendix I, https://links.lww.com/IJEBH/A118).

Participants

District nurses (European Qualifications Framework [EQF-5/6]), nurses (EQF-4), and nurse assistants (EQF-2/3) working in home care were sampled purposively from one large home care organization in the south-west Netherlands. This organization employed approximately 800 home care nurses at the time of the study and had used the Medido dispenser since 2013.

In sampling, we ensured maximum variation in geographic areas of operation (by including various work cultures and team processes, and neighborhoods where the nurses work), educational level (EQF),27 age, sex, years of work experience, and level of experience with AHMD.

The main inclusion criterion was working in a team that provides regular (non-specialized) home care. Seconded employees and professionals with insufficient mastery of the Dutch or English language were excluded.

Procedures

Managers of the home care organization nominated eligible nurses. We selected participants based on the maximum variation criteria by asking each manager, representing different geographic areas, to nominate two nurse assistants EQF-2, two nurse assistants EQF-3, two nurses EQF-4, and two district nurses EQF-5/6 (including information on age and years of work experience). From there, we selected two or three names per manager, with as much variation as possible in all factors. Potential candidates received a letter with information about the subject, the aim of the study, and the study procedures. After we received written informed consent, interviews were scheduled. Interviews took place in March and April 2021 and were conducted by CM. Due to COVID-19 restrictions, online videoconferencing software (Microsoft Teams, Microsoft, Redmond, Washington, USA) was used. The interviews were audio-recorded and video-recorded, transcribed verbatim, and pseudonymized. Sampling continued until no new patterns or themes emerged (thematic data saturation).28 Then, two more interviews were conducted to ensure data saturation was reached.

Data collection

The seven domains of the Tailored Implementation for Chronic Diseases (TICD) checklist29 were used to derive a comprehensive overview: (1) guideline factors; (2) individual health professional factors; (3) patient factors; (4) professional interactions; (5) incentives and resources; (6) capacity for organizational change; and (7) social, political and legal factors.29 Because AHMD is an innovation rather than a guideline, the first domain was supplemented with the innovation factors of Grol and Wensing30: advantages in practice, feasibility, credibility, accessibility, and attractiveness. These domains were translated into an interview guide that fitted daily practice (Appendices II, https://links.lww.com/IJEBH/A119 and III, https://links.lww.com/IJEBH/A120). This was piloted in one interview, as described by Creswell and Poth.31 No further changes were deemed necessary. This interview was included in the main analysis.

Data analysis

Thematic content analysis started after the first interview. New insights gained from consecutive analyses were included in the successive interviews, and constant comparison was applied until data saturation.32,33

The first four transcripts were initially open-coded inductively to ensure important themes were not lost through deductive analysis.34 After that, a code tree was developed, based on the TICD checklist29 and the inductive codes (Appendix IV, https://links.lww.com/IJEBH/A121). The code tree was constantly adapted during data analysis. When data were relevant to multiple codes, they were included in all relevant codes. Using the method of Dierckx de Casterle et al.,35 we created overviews of the codes/determinants per interview labelled as barrier, facilitator, or both (conceptual interview scheme35). Determinants were labelled as “both” when on the one hand it was a facilitator, but on the other hand it was also a barrier. Finally, all determinants were combined into a comprehensive overview of barriers and facilitators. To determine which domains were dominant, we counted the determinants (derived from our code tree) per TICD domain. In addition, we developed a checklist of practical strategies to guide the implementation of AHMD in home care (Appendix V, https://links.lww.com/IJEBH/A132). Data were managed and analyzed in NVivo 12 (Lumivero, Denver, Colorado, USA).

Study rigor

Member checking during the interviews was performed by probing the respondents to ensure that their opinions were correctly understood. After analysis, written feedback was requested from the respondents on the conceptual interview schemes.31 Through peer reviewing, the reliability of the coding was strengthened.33 EI and an independent researcher coded two different, randomly chosen interviews, and discrepancies were discussed until consensus was reached. Member checks and peer reviews were performed to strengthen the credibility of our study. Memos about context, observations, and methodological choices were written after each interview, thereby creating an audit trail, useful to judge validity and to strengthen the study's dependability and confirmability33 (Appendix VI, https://links.lww.com/IJEBH/A123).

Ethics

The Medical Ethics Research Committee of the Erasmus University Medical Center Rotterdam approved this study (MEC-2021-0071). Written informed consent was obtained prior to all interviews. Respondent confidentiality was ensured by pseudonymizing the transcripts, and only members of the research team had data access. The study conforms to the Declaration of Helsinki (as revised in Tokyo in 2004).

RESULTS Respondents

Eighteen nurses were invited. Two declined participation for personal reasons, and one did not respond. Thus, 15 nurses participated. Non-participants were not present. Ages ranged from 23 to 65 years (median 34 years). Years of work experience varied from 2 to 48 years (median 10 years). All respondent characteristics are presented in Table 1.

Table 1 - Characteristics of the interviewed home care nurses Characteristic Participants (n = 15) Sex (female), n 12 Age (years), median 34 Educational level (EQFa), n  EQF-2 2  EQF-3 1  EQF-4 7  EQF-5/6 (district nurse) 5 Years of work experienceb, n  2–9 years 6  10–29 years 6  30–48 years 3 Number of patients with AHMD, n  1–5 patients 10  6–10 patients 4  >10 patients 1

AHMD, automated home medication dispenser; EQF, European Qualifications Framework.

aEuropean Qualifications Framework25 for education level.

bYears of work experience was measured as a continuous variable, and then classified into a categorical variable to ensure the anonymity of the respondents.

The mean duration of the interviews was 50 minutes (range 39−72 minutes). Thematic data saturation was reached after 15 interviews. Thirteen respondents returned member checks. Four provided additional information, which was included in the analysis. The peer reviews did not reveal major discrepancies and new determinants did not come to the fore.

Seventy-eight determinants (barrier, facilitator, or both) were identified (Table 2 and Table 3). Sixty-four (82%) fell within the first three TICD domains29: innovation factors (n = 16), individual health professional factors (n = 20), patient factors (n = 28), professional interactions (n = 6), incentives and resources (n = 2), and capacity for organizational change (n = 6). Determinants in the social, political, and legal factors domain were not identified. Therefore, this domain is not further reported in this article. A comprehensive overview of the barriers and facilitators is presented in Table 4.

Table 2 - Domains of the Tailored Implementation for Chronic Diseases checklist and the numbers of determinants coded within the domain TICD domain Barriers, n Facilitators, n Both, n Total, n Innovation factors 3 6 7 16 Individual health professional factors 3 6 11 20 Patient factors 4 12 12 28 Professional interactions 1 – 5 6 Incentives and resources – – 2 2 Capacity for organizational change 1 4 1 6 Social, political, and legal factors – – – – Total, n 13 28 37 78

TICD, Tailored Implementation for Chronic Diseases.

Table 3 - Determinants for the implementation of automated home medication dispensers labelled as barrier, facilitator, or both, and the number of mentions TICD domain Determinant Barrier Facilitator Mentioned by x of the 15 respondents, n Number of times mentioned in total, n Innovation factors Disadvantages in practice  Alarms + unplanned visits X 14 62  Nurses having less control over the situation X 2 5  Less social contact X 7 8 Advantages in practice  Creative solutions to prevent and fix minor errors X 4 10  Quality of life X 4 4  Quality of treatment X 9 17  Sachets are cut open X 4 4  Self-sustainability X 13 33 Accessibility X X 10 14  Home care is required X X 3 7 Attractiveness X X 3 4 Conditions X X 7 18 Credibility X 13 23 Efficiency in home health care X X 12 27 Feasibility X X 5 7 Not portable (but early take-out possibility) X X 5 7 Individual health professional factors Cognition (including attitudes)  Feeling competent and confident X X 11 30  Feeling frustrated X 4 9  Intention and motivation X X 9 18  Perspectives   Perspectives on AHMD X X 7 13   Perspectives on eHealth (in general) X X 12 23 Knowledge and skills X X 7 12  Awareness and familiarity X X 13 36  Clarity X 1 2  Administration X 6 7 Make it simple for home care professionals X X 1 3 Point of contact for asking questions X 3 3 Products and tools for home care professionals X 11 32 Sustainability X X 7 13 Training X 10 31  AHMD dummy for practice X 13 25  Coaching on the job X 7 20  E-learning X X 10 25  Learning from others’ experiences X 4 8 Professional behavior X X 1 2  (Impact of) past experiences X X 8 21 Patient factors Anxiety or panic X 8 10 Concerns X 10 11  Family and informal caregivers X X 5 12  Family's opinions and experiences X X 7 13 Involving family X X 8 17 Needs and wishes X 2 2  (Phasing out) guidance X 11 31  Evaluation X 4 12  Feeling confident X X 3 3  Keep or make it simple (patient) X 5 19  Knowledge and instructions X 12 28   Products and tools (patient) X 10 22   YouTube videos X 6 15  Point of contact for questions X 6 9 Opinions and experiences X X 9 22 Patient motivation X X 7 25  Explaining advantages X 12 19  Relationship of trust X 7 11  Trying without obligations X 9 17 Selecting patients X X 12 24  Being away from home X X 4 11  Combination with other care X X 2 3  Forgetting medication X X 8 13  Knowing the patient, tailored decision X 3 4  Learnability X X 6 10   Cognitive impairment X 14 36  Non-adherent on purpose X 4 12  Patient selection tool X X 8 12 Professional interactions Collaboration and support  External collaboration X X 8 16  Internal collaboration X X 14 21 Communication X X 4 4 Involving whole team (signalling role) X X 8 18 Helpdesk employees available by phone call to adjust settings X X 8 12 Takes too much time to reach the helpdesk; some nurses wish to be able to adjust settings themselves, without needing the helpdesk X 3 10 Incentives and resources Financial (dis)incentives X X 12 29 Quality and safety assurance X X 14 33 Capacity for organizational change Leadership X 1 1  Direction from manager X 3 4