Table 1 Characteristics of included papers examining family involvement for managing medications of older patients across transitions of care (N = 23)

Allen et al. 2017 [13] Australia***

Design: Qualitative exploratory.

Setting: metropolitan public health-care network

Data collection: interviews.

Transition points involved: acute ward setting and rehabilitation setting in hospital, home.

Medication management process involved: admission instruction, discharge instruction.

Patients n = 19.

78.9 years (age range 45–94 years for patients and families).

Gender: not stated.

Family n = 7.

Gender: not stated.

Caring relationships with health professionals:

- Nurses in rehabilitation attended to follow-up phone calls to check on discharge management.

- Lack of continuity of medical practitioners and interactions with multiple medical practitioners.

- Medical decision making about discharge medications without understanding about medications prescribed by other medical practitioners.

Seeking information:

- If patients was too unwell to seek information during their acute illness, family wanted to have medication information on their behalf.

- Expectation that doctors would share medication information with them during the hospital admission. This did not always occur.

- Family members were well informed about changes to medications in rehabilitation ward setting.

- General practitioner (GP) relied on accurate and timely discharge summary to explain medication information to family.

Coleman et al. 2006 [15] United States**

Design: Randomised controlled trial.

Setting: one hospital, eight skilled nursing facilities, one home health care agency.

Data collection: rates of rehospitalisation measured at 30, 90, and 180 days.

Transition points involved: community; skilled nursing facility; rehospitalisation.

Medication management processes involved: medication self-management.

Intervention – four pillars.

- Support patients and family with medication self-management, medication reconciliation.

- Patient-centred record to assist with site transitions.

- Timely follow-up with care.

- Supply patients and family with list of “red flags” for worsening condition.

Patients:

n = 379 intervention group,

n = 371 control group.

Age for intervention group: mean 76.0 years (SD 7.1 years).

Age for control group: mean 76.4 years (SD 6.8 years).

Gender: 48% female for intervention group.

52% female for control group.

Medical condition: at least 1 of 11 selected acute or chronic conditions.

Family recruited with patients:

n = not specified.

Relationship with patient: not specified.

Outcomes:

Rehospitalisation rates at 30 days –

Intervention group = 8.3%

Control group = 11.9%, p = 0.048.

Rehospitalisation rates at 90 days –

Intervention group = 16.7%

Control group = 22.5%, p = 0.040.

Rehospitalisation rates for the same condition that precipitated the index hospitalisation at 90 days –

Intervention group = 5.3%

Control group = 9.8%, p = 0.04.

Rehospitalisation rates for the same condition that precipitated the index hospitalisation at 180 days -

Intervention patients: 8.6%

Control patients: 13.9%, p = 0.046.

Mean hospital costs at 180 days -

Intervention patients: $2058

Control patients: $2546, log-transformed p = 0.049.

Coleman et al. 2004 [14] United States***

Design: Quasi experimental study with intervention and control groups.

Setting: one community, hospital.

Data collection: rates of rehospitalisation at 30, 90 and 180 days; care transition measure

Transition points involved: hospital, community residence facility, home.

Medication management process involved: medication self-management.

Intervention – four pillars.

- Support patients and family with medication self-management, medication reconciliation.

- Patient-centred record to assist with site transitions.

- Timely follow-up with care.

- Supply patients and family with list of “red flags” for worsening condition.

Intervention involved: meetings with transitions coach; communication tool (personalised medical record, role plays), follow up phone calls.

Patients:

n = 158 intervention group,

n = 1235 control group from administrative data.

Age for intervention group: mean 75.1 years (SD 6.4 years).

Age for control group: mean 78.5 years (SD 7.5 years).

54% female for intervention group.

55% female for control group.

Medical condition: at least one or more of 9 medical conditions.

Family recruited with patients:

n = not specified.

Relationship with patient: not specified.

Outcomes:

Rehospitalisation at 30 days –

Intervention group compared with control group: 0.52 (95% confidence interval (CI) 0.25, 0.96)

Rehospitalisation at 90 days –

Intervention group compared with control group: 0.57 (95% CI 0.25, 0.72)

Rehospitalisation at 180 days –

Intervention group compared with control group: 0.57 (95% CI 0.36, 0.92).

Intervention group - patients reported: high level of confidence in obtaining information, communicating with health care team, and understanding their medication regimen.

Quality of care transitions taxonomy –

Intervention group: 9.5% of posthospital transitions were complicated.

Control group: 14.9% of transitions, P = 0.35.

Crawford et al. 2015 [16] Australia***

Design: Qualitative, exploratory.

Setting: one aged rehabilitation and geriatric evaluation and management facility

Data collection:

Semi-structured interviews.

Transition points involved: acute hospital setting, aged rehabilitation and geriatric evaluation and management facility

Medication management process involved: discharge planning.

Family n = 20.

Relationship with patient: husbands, wives, daughters, son, daughter-in-law, close friends of people with dementia.

Themes:

Adjusting to new role from caregiver to visitor –

- Difficulty in relinquishing role.

- Desire to continue to be involved in decisions.

- Felt ongoing responsibility to communicate medication needs.

- Belief about specialised knowledge about relative’s care.

- Staying informed helped caregivers to cope with move to facility.

Deeks et al. (2016) [17] Australia**

Design: Qualitative, exploratory.

Setting: three urban primary and acute sites, one rural primary and acute site.

Data collection: semi-structured interviews.

Medication management process involved: hospital admission and discharge for patients with dementia.

51 participants comprising doctors, nurses, pharmacists, occupational therapists, general practitioners, Alzheimer’s Australia staff, and family.

Family n = not specified.

Relationship with patient: not specified.

Themes:

Medication reconciliation –

- Verifying accurate list on admission was difficult if family members not present or medications not brought to hospital.

Lack of modified planning for care transitions –

- Lack of identification that patient had cognitive problems as they moved across settings.

- Inadequate information about medication changes at discharge.

- Carers’ support for use of dose administration aids by patients, but potential problems with errors. Lack of assessment of patients’ ability to use aid.

- Desire for once-daily dosing if paid carer required to visit at home after discharge.

Multiple prescribers –

- Little information shared between private and public hospitals.

- Treatment delays between specialist and general practitioner as communication by letter.

Residential aged care facilities –

- Lack of accurate and complete information from hospital.

Medication reviews by pharmacists –

- Patients with dementia trusted and built rapport with community pharmacists and general practitioners. Patients not receptive to home medicine reviews.

Dyrstad et al. (2015) [18] Norway**

Design: Qualitative, exploratory.

Setting: two hospitals

Data collection: observations, conversations, observational field notes.

Transition points involved: two emergency departments (EDs), seven hospital wards; including admissions from home based care or nursing homes, and discharge.

Medication management process involved: admission instruction, discharge instruction.

Patients n = 41.

Age: mean 86.0 years, range 73–97 years.

Gender: 46% female.

Medical condition: an orthopaedic diagnosis or chronic condition, and poly-pharmacy (> 5 medications daily).

Family n = 28.

Relationship with patient: son, daughter, wife.

Health professionals n = not specified.

Disciplines involved: paramedics, nurses, doctors.

Themes:

Information dissemination and decision-making –

- No scheduled discharge planning meetings with patient and family.

- Nurse phoned family to inform them of doctors’ decisions.

Next of kin important advocates –

- In ED, provided valuable information about medications taken before admission.

- Family administered medications in ED during busy times.

- In wards, no routines to invite family to participate on doctor’s rounds.

- Informed on day of discharge about ward round decisions.

- Some families had to seek information about medication decisions at discharge.

Georgiadis & Corrigan 2017 [19] United Kingdom***

Design: Phenomenological.

Setting: three hospitals.

Data collection: Interview, audio diary, written diary.

Transition points involved: clinical settings (not described), home.

Medication management process involved: medication counselling at discharge.

18 participants

Patients n = 12.

Age: 65.9 years (SD 17.2 years).

Gender: Mixed, not stated.

Medical condition: not mentioned. Patients with non-medical complex conditions, which were not defined.

Family n = 6.

Relationship with patient: not specified.

Themes:

Limited involvement in discharge-care preparations –

- Premature discharges meant lack of planning in medication counselling.

- Unexpected and delayed discharges meant that family unable to be present.

Weak service interface –

- Expected organisation of appointments for outpatient clinic or home visits did not happen. Reliance on family for medication administration.

- Family arranged primary clinic to assess medications.

Hagedoorn et al. 2017 [20] The Netherlands***

Design: Qualitative exploratory.

Setting: Four general hospitals,

Data collection: non-participant observations, audio-recordings of planned discussions for admission and discharge discussions and family meetings.

Transition points involved: admission to ward, 13 clinical wards comprising neurology, pulmonary, internal medicine, cardiology, geriatrics. Discharge from ward.

Medication management process involved: administration, monitoring.

Patients n = 62.

Age: 76 years (SD 7.2 years).

Gender: 48% female.

Medical condition: 22 patients (36%) had 3 or more chronic diseases.

Family n = 107 at planned discussions.

Relationship with patient: husband, wife, son, daughter.

Themes:

Social network support –

- Support by family caregivers assisting with or monitoring medication intake at home. Not addressed by nurses in hospital assessment sessions.

Coordination of care –

- During discharge discussions nurses reviewed home medication list with patients and family.

- Family asked specific questions about changes in the patients’ home medications.

Hvalvik & Reierson, 2015 [21] Norway***

Design: Phenomenological hermeneutic design.

Setting: hospital.

Data collection: in-depth interviews.

Transition points involved: hospital, municipal rehabilitation, short-term care facility, home.

Medication management process involved: discharge planning, self-management at home.

Family n = 11.

Relationship with patient: son, spouse, daughter.

Themes:

Balancing vulnerability and strength –

- Enduring emotional stress with discharge.

- Family expected to be involved in discharge, but were not included.

- Worry about being discharged too early and lack of medication information.

- Lack of communication about medications during stay.

- Sense of responsibility in managing medications.

- Lack of communication between hospital and community services. Prescription of medications that should have been ceased due to adverse effects or allergies.

Coping with an altered everyday life –

- Dealing with changes to family routines.

- Anticipating possible problems if patients discharged early.

- Comprehensive understanding about older person’s vulnerability, and fragility.

Jeffs et al. 2017 Canada***

Design: Qualitative exploratory.

Setting: orthopedic inpatient units in two acute care hospitals and one orthopedic unit at a complex continuing care rehabilitation hospital.

Data collection: semi-structured interviews.

Medication management process involved: care involving transfer from an acute care hospital to a rehabilitation hospital.

Patients n = 13.

Age: 82.9 years (range: 68–91 years).

Gender: 69% female.

Medical condition: non-elective patients who had fallen or sustained a fracture though an accident.

Family n = 9.

Relationship with patient: child, spouse, partner, sibling.

Health professionals n = 50.

Themes:

Watching –

- Alert to medication administration and changes in patients’ status.

- Patients’ lack of understanding about medication changes.

Being an active care provider –

- Responsibility for providing care that health care providers performed.

- Administering medication was easier for family to do than the nurses, as nurses experienced challenges with the patient taking medication.

- Lack of engagement by health care providers about involving family.

Advocating –

- Being supportive of patients’ needs such as changing medication times to suit patients’ routines.

Navigating the health care system –

- Asking questions and coordinating follow-up care.

- Arranging appointments with various members of interdisciplinary team.

- Lack of availability of health care providers to ask questions about transitions plan.

King et al. 2013 [23] United States***

Design: Qualitative study using grounded dimensional

Analysis.

Setting: five non-profit religious and government skilled nursing facilities (SNF)

Data collection: focus group and individual interviews

Transition points involved: skilled nursing facilities, hospital.

Medication management process involved: hospital discharge.

Health professionals n = 27.

Disciplines involved: nurses.

Themes:

Reconciling hospital information-

- Seeking medication details from families was problematic as sometimes not adequately informed about family members contact details.

- Asking families about medication details created a poor first impression of SNF staff.

Consequences of poor-quality discharge communication –

- Care delays and implementation of an inappropriate medication plan.

- Inaccurate hospital information produced family dissatisfaction.

- Made the SNF appear unorganised.

Knight et al. 2013 [24] United Kingdom**

Design: Qualitative exploratory.

Setting: participants’ home

Data collection: semi-structured interviews, medication diary.

Transition points involved: hospital, home

Medication management process involved: discharge process.

Patients n = 7.

Age: > 75 years.

Gender: 43% female.

Medical condition: not stated.

Family n = 12.

Relationship with patient: wife, husband.

Themes:

Discharge in general –

- Long delays till discharge or abrupt notification about discharge.

Obtaining medication for discharge-

- Waiting for medications to be prepared at hospital pharmacy.

Information regarding discharge medication –

- Carer belief that it was their responsibility to check understanding on individual medicines.

- Carer satisfaction about information provided but not detailed.

- Medication changes in hospital not conveyed to carers.

- Inadequate explanations of new medicines and associated risks for the patient.

Medication lists –

- Lack of written guide available or had never received a list.

Communication about medication in hospital and following discharge –

- Carers detected medication omissions after careful examination.

- Needed to feel better prepared with medications post-discharge.

Lowson et al. 2012 United Kingdom***

Design: Qualitative exploratory.

Setting: participants’ home

Data collection: semi-structured interviews.

Transition points involved: hospital, home.

Medication management process involved: medication information at admission.

Patients n = 27.

Age: mean = 79.0 years (SD 4.25 years).

Gender: 52% female.

Medical condition: heart failure or cancer in the last year of life.

Family n = 12.

Relationship with patient: wife, husband, daughter, sister, neighbour, friend.

Themes:

Conductors –

- Strong contributions to maintaining good care throughout illness trajectory.

- Detailed knowledge about medications.

- Alerted health professionals about potential medication errors.

Second fiddle –

- Following hospital admission, ability to work with health professionals to influence decisions vastly reduced.

- Carers invested effort in maintaining continuity of relationship.

- Advocated on patients’ behalf to affect beneficial change.

Nazarath et al. 2001 United Kingdom***

Design: Randomised controlled trial.

Setting: three general hospitals, one long-stay hospital

Data collection:

Transition points involved: general hospital wards, home.

Medication management process involved: discharge information.

Intervention: (for intervention studies) Discharge plans developed by pharmacists, home visit by community pharmacist, counselling patients and family on appropriate doses and purpose.

Control group: discharge letter to general practitioner.

Patients:

n = 181 intervention group,

n = 181 control group.

Age for intervention group: mean 84 years (SD 5.2 years).

Age for control group: mean 84 years (SD 5.4 years).

Gender:

62% female for intervention group.

66% female for control group.

Medical condition: had a mean of three chronic conditions.

Family (n = not specified)

Included in intervention but not stated.

Outcomes:

Hospital readmission at 3 months –

Intervention group: 64 (39%)

Control group: 69 (39.2%), p > 0.05.

Hospital readmission at 6 months –

Intervention group: 38 (27.9%)

Control group: 43 (28.4%), p > 0.05

No differences between groups: Patients’ general well-being, satisfaction with the service and knowledge of and adherence to prescribed medication (p > 0.05).

Neiterman et al. 2015 [25] Canada**

Design: Qualitative exploratory.

Setting: participants’ home

Data collection: interviews

Transition points involved: hospital, home

Medication management process involved: medication management at home

Patients n = 17.

Age: 70–89 years, mean = 79 years.

Gender: 41% female.

Medical condition: diverse chronic illnesses.

Family n = 19.

Relationship with patient: husband, wife, mother, father, daughter, son, daughter-in-law, son-in-law.

Themes:

Dealing with medical confusion –

- Post-discharge medication management was difficult because of medication changes.

Facilitators for recovery: social capital and social support -

- Family members made sure that medications schedules were followed.

- Family was overwhelmed and exhausted due to constant need to coordinate care and ensure patients’ needs met.

Targeted nurse practitioner initiative -

- Some caregivers did not fully understand the nurse practitioner (NP) role but relied on NPs to oversee the management of medications.

Palagyi et al. 2016 [26] Australia***

Design: Qualitative exploratory.

Setting: three long-term care facilities

Data collection: Focus groups and interviews

Transition points involved: long-term care facilities, hospital, community care.

Medication management process involved: deprescribing medications.

Patients n = 25.

Age: mean = 87.6 years, range 75–100.

Gender: 77% female.

Medical condition: not stated.

Family n = 16.

Relationship with patient: not mentioned.

Health professionals n = 27.

Disciplines involved: general practitioners, long-term care facilities staff.

Themes:

Pitfalls of coordinated care -

- Lack of review of acute medication after condition was treated.

Negotiating a complex system -

- Family concerned that compulsory two-year residential medication management review schedule was too long. Many changes occur in two years.

Medication knowledge -

- Family were unfamiliar with specific medication indications.-

Minimal recognition of adverse drug reactions.

Whatever the GP says goes -

- Complete trust in the care and decisions of the GP.

- Number of specialists involved in residents’ care.

Need for realistic expectations-

- Relatives viewed long-term care facilities as active providers of medical care. GPs regarded it as a palliative care environment.

Ploeg et al. 2017 [31] Canada***

Design: Interpretive descriptive.

Setting: participants’ home

Data collection: semi-structured interviews.

Transition points involved: home, primary care settings, hospital.

Medication management process involved: managing medications for multiple comorbidities.

Patients n = 41.

Age: 17% aged 85 years and over.

Gender: 44% female.

Medical condition: three or more chronic conditions.

Family n = 47.

Relationship with patient: wife, husband, grandfather, mother-in-law, friends.

Health professionals n = 42.

Disciplines involved: registered nurse, registered/licensed practical nurse, personal support worker/healthcare aide

Themes:

Experience of managing multiple chronic conditions -

- Emotionally draining.

Organising pills and appointments -

- Managing changes to medications that frequently occurred after an acute care hospitalization.

- Abrupt discharge, without input from caregivers about preferences.

- Organising appointments to discuss blood results and scans affecting medications.

Being split –

- Receiving services from multiple providers who focus on a single disease

- Lack of communication between family doctor and specialists.

Doing what the doctor says -

- Family believed decision making was physician-directed. Providers perceived their approach involved shared decision-making.

Popejoy 2011 [32] United States***

Design: Qualitative exploratory.

Setting: tertiary care hospital

Data collection: semi-structured interviews.

Transition points involved: hospital, home, residential aged care facility.

Medication management process involved: Discharge planning.

Patients n = 13.

Age: mean = 84 years, range = 72–89 years.

Gender: 62% female.

Medical condition: no cognitive impairment, at least pre-clinically frail.

Family n = 12.

Relationship with patient: spouses.

Health professionals n = 7.

Disciplines involved: registered nurses, social workers.

Themes:

Changing the discharge plan -

- Patients and caregivers were adamant about going home (not to residential care).

- Health professionals were worried when family had trouble understanding about medications.

- Family were sometimes old with health problems themselves, and had difficulties coping with and remembering to offer patients’ medications at home.

Tjia et al. 2014 [27] United States***

Design: qualitative exploratory.

Setting: Three hospice agencies.

Data collection: semi-structured interviews

Transition points involved: hospice, outpatient oncology, primary care settings.

Medication management process involved: medication prescribing across transitions.

Patients n = 18.

Age: mean = 80 years (SD 10 years).

Gender: 42% female.

Medical condition: advanced cancer.

Family n = 8.

Relationship with patient: not stated.

Health professionals n = 17.

Disciplines involved: nurses, physicians.

Themes:

Medication coordination and communication -

- Families were keen to have comprehensive medication reviews upon transition to hospice that assessed ongoing use of longstanding medications for comorbid illness.

- Family were receptive to reducing harmful and non-essential medications.

Towle et al. 2012 [12] Singapore*

Design: Prospective observational, cross-sectional.

Setting: one tertiary hospital

Data collection: survey questionnaire

Transition points involved: hospital, home.

Medication management process involved: Discharge process.

Intervention: (for intervention studies) BOOST (Better Outcome for Older adults through Safe Transitions); an evidence-based quality improvement initiative to enhance care transition in improving patient/family preparedness for discharge.

Patients n = 40.

Age: not stated.

Gender: not stated.

Medical condition: > 1 chronic condition.

Family n = not stated.

Relationship with patient: not stated.

Outcomes:

Patient and caregiver understanding of medical condition -

- Improvement by 70%.

Patient and caregiver understanding of medications-

- Improvement by 67%.

Patient and caregiver understanding of treatment plan -

- Improvement by 81%.Patient and caregiver understanding of follow-up -

- Improvement by 41%.

Trollor 1997 [28] Australia****

Design: Cross-sectional.

Setting: Community palliative care service.

Data collection: questionnaire.

Transition points involved: palliative care services, home, primary care services.

Medication management process involved: symptom management.

Patients n = 26.

Age: not stated.

Gender: 31% female.

Medical condition: patients with palliative care needs.

Family n = 26.

Relationship with patient: wives, daughters, husbands.

Themes:

- 13 out of 26 family members were in charge of patients’ medication.

- Role in dealing with general practitioners and palliative care specialists, to administer medication for pain, sleeping difficulty and loss of appetite were most challenging.

White et al. 2015 [29] United States***

Design: Qualitative exploratory.

Setting: Two hospitals.

Data collection: semi-structured interviews.

Transition points involved: hospital, home.

Medication management process involved: Discharge planning.

Patients n = 20.

Age: mean = 77 years (SD 8.8 years).

Gender: 47% female.

Medical condition: stroke survivors.

Family n = 9.

Relationship with patient: husband, wife, son, daughter.

Themes:

Preparing to go home after the stroke -

- Importance of health professionals understanding family’s needs following discharge so that specific situation could be addressed.

- Wrong family member targeted for providing information.

Complexity of medication management-

- some family were confident about their knowledge.

- Some family lacked understanding about the purpose of medication.

- Usually family member who managed medications.

- Difficulties in managing medications for patients with swallowing problems. Random decisions about which medications should be crushed.

White et al. 2014 [33] United States**

Design: Mixed methods.

Setting: Two hospitals.

Data collection: semi-structured interviews, electronic medical record review.

Transition points involved: hospital, home.

Medication management process involved: Discharge planning.

Patients n = 310.

Age: mean = 76 years (SD 9.8 years).

Gender: Not stated.

Medical condition: stroke survivors.

Family n = 20 combined with patients.

Relationship with patient: not stated.

Themes:

Within one month of discharge, 10% were readmitted and 25% within 6 months.

Reasons for readmission were recurrent stroke/transient ischaemic attack (19%), pneumonia and urinary tract infection (19%), swallowing problems and dehydration (9%), and cardiac causes (7%).

- Need for guidance on what to expect at home.

- Need follow-up in community about early identification of problems.

- Complexity of medication management sometimes led to lack of understanding.