Article Text
PDF
Abstract
Objectives The study examines whether hospital discharge practices and care-transition preparedness mediate the association between patients’ cultural factors and readmissions.
Methods A prospective study of internal medicine patients (n=599) examining a culturally diverse cohort, at a tertiary medical centre in Israel. The in-hospital baseline questionnaire included sociodemographic, cultural factors (Multidimensional Health Locus of Control, family collectivism, health literacy and minority status) and physical, mental and functional health status. A follow-up telephone survey assessed hospital discharge practices: use of the teach-back method, providers’ cultural competence, at-discharge language concordance and caregiver presence and care-transition preparedness using the care transition measure (CTM). Clinical and administrative data, including 30-day readmissions to any hospital, were retrieved from the healthcare organisation’s data warehouse. Multiple mediation was tested using Hayes’s PROCESS procedure, model 80.
Results A total of 101 patients (17%) were readmitted within 30 days. Multiple logistic regressions indicated that all cultural factors, except for minority status, were associated with 30-day readmission when no mediators were included (p<0.05). Multiple mediation analysis indicated significant indirect effects of the cultural factors on readmission through the hospital discharge practices and CTM. Finally, when the mediators were included, strong direct and indirect effects between minority status and readmission were found (B coefficient=−0.95; p=0.021).
Conclusions The results show that the association between patients’ cultural factors and 30-day readmission is mediated by the hospital discharge practices and care transition. Providing high-quality discharge planning tailored to patients’ cultural characteristics is associated with better care-transition preparedness, which, in turn, is associated with reduced 30-day readmissions.
- cultural and linguistic factors
- minorities
- discharge practices
- care-transition preparedness
- 30-day readmission
Statistics from Altmetric.com
- cultural and linguistic factors
- minorities
- discharge practices
- care-transition preparedness
- 30-day readmission
Introduction
The growing international focus on readmission reduction highlights the need to assess the quality of transitional care and discharge practices.1–4 Recent efforts to better understand causes of readmission and how to prevent them have shifted from a provider-centric approach to a patient-centric approach, focusing on patients’ experience and perspectives of their transition from hospital to home.5–7 A positive patient experience during the transition process, which centres on discharge preparedness and patients’ understanding of medications and follow-up instructions, has been linked to reduced readmissions.8–11 For example, a study on patients’ perceptions of the care they receive during the index hospitalisation 12 showed that patients who report a positive, engaging interpersonal-care experience have lower readmission rates.
Achieving a positive care-transition experience is dependent on the quality of hospital discharge practices. Intervention studies have shown that when patients engage in open ongoing communication with their providers, including the ability to answer questions and confirm understanding, this affects their overall care-transition experience.13 14 Studies that specifically examined physician–patient communication during the index admission15 and the provision of extensive discharge counselling16 showed a significant positive association with overall care-transition preparedness.
An especially vulnerable group, in need of a personally tailored care transition approach, is that of ethnic minorities. Patients from minority groups are at greater risk for care-transition breakdowns because of cultural and linguistic barriers that affect the comprehension of medical instructions and the ability to navigate self-care among the various settings.16 17 Evidence points to significant disparities in readmission according to racial or ethnic origins.17–19 Yet, a broad understanding of how culture, that is, a system of shared values and beliefs reflecting patients’ social environment, lifestyle attitudes and proficiencies, affects patients’ behaviours and decision-making, as well as the responsiveness of the healthcare system to patients’ needs, is lacking.20 21
This study aims to fill this gap by examining whether patients’ minority status (ie, whether or not they belong to an ethnic minority group), their health-related locus of control, family collectivism and health literacy are associated with hospital discharge practices (ie, the degree to which the discharge process is tailored to address challenges of patients from diverse cultural backgrounds), patients’ overall care-transition preparedness (ie, their perceptions that they know how to manage their posthospital care) and outcomes (readmissions). We build on knowledge accumulated in two previous studies performed as part of a larger project studying transitional care of diverse population groups. We previously showed that patient-provider language-concordant care and caregiver presence at-discharge are associated with better care-transition preparedness among patients with low health literacy levels and among minorities in general.22 Additionally, we showed that the extent of explanations provided at the in-hospital discharge briefing and during the postdischarge primary care visit are significantly associated with reduced readmissions.23 In this study, we aim to add to previous knowledge by testing the mediating role of hospital discharge practices and care-transition preparedness in the association between patients’ diverse cultural factors and readmissions.
Methods
Study setting and participants
We conducted this prospective cohort study between June 2013 and July 2014, as part of a lager study that examined cultural factors, transitional care processes and outcomes.22 23 The study took place at a large medical centre in Israel, following patients hospitalised in six internal-medicine wards, and up to 30 days after discharge. Israel provides a good setting to study transitions of minorities as it includes several ethnic minority groups, the largest are native Russian-speaking Jewish immigrants from the former Soviet Union and Arabic-speaking Muslims, Christians and Druze .24 This diversity is also reflected in the multicultural workforce of the Israeli healthcare system, which often enables the provision of culturally and linguistically concordant care.25 26 Moreover, the culturally and linguistically diverse workforce is increasingly trained in tailoring care to meet the needs of the wide range of ethnic and cultural patient populations.27 Patients were included if they were members of Clalit Health Services, the largest not-for-profit integrated healthcare provider and insurer in Israel, age 18 years or older, hospitalised for an unplanned admission of at least one night and self-defined native Hebrew, Arabic or Russian speaking. Previous Israeli studies have shown that Arabic and Russian speakers (recent immigrants from the former Soviet Union) comprise about 11% and 37% (respectively) of hospitalised older adults.28 29 We oversampled minority Arabic speakers to enable a large enough sample for testing additional cultural factors (such as health locus of control and family collectivism). Patients were not included if they had a diagnosis of cognitive failure, received palliative or end-of-life care and/or lacked an active telephone number for follow-up.
Data collection
In-hospital baseline questionnaires were administered in the patients’ self-identified native or main spoken and read language (Hebrew, Arabic or Russian) to collect patients’ sociodemographic, cultural and linguistic characteristics and physical, mental and functional health status. Three to 7 days after discharge, patients were surveyed by phone about the in-hospital discharge practices they experienced and about their care-transition preparedness. Information about participants’ chronic conditions, length of stay, prior hospitalisation and 30-day readmissions to any hospital was retrieved from Clalit’s electronic health records and administrative data warehouse. The study was powered to detect differences in 30-day readmission rates with the application of G*Power analysis for two independent means30 based on estimates of a 15%–17% readmission rate,31 with a power of 80% and alpha of 0.05.
Measures
Outcome
The main outcome was readmission, defined as any unplanned hospitalisation (to any general hospital in Israel), within 30 days of discharge from the index hospitalisation.
Predictors
Health locus of control was measured using three subscales of Walston’s Multidimensional Measure of Health Locus of Control (HLC) form B32 33: internal HLC, people’s beliefs that they are responsible for their own health outcomes; powerful others HLC, the belief that healthcare professionals and other people are responsible for one’s health outcomes and chance HLC, the perception that luck/random effects are the most powerful predictors of health outcomes. Each subscale is measured by six items rated from 1=absolutely disagree to 6=absolutely agree. The Cronbach’s alpha coefficients for the internal HLC, powerful others HLC and chance HLC were 0.92, 0.90 and 0.85, respectively.
Family collectivism was measured with the self-reported six-item collectivism subscale of the cultural constructs developed by Lukwago et al.34 The scale reflects an emphasis on family obligation and belief in family importance. Answers are rated on a four-point response scale (1=not at all important, 4=very important) (Cronbach’s alpha=0.88).
Health literacy was assessed using the Brief Health Literacy Screen (BHLS), a seven-item questionnaire answered on a five-point Likert-type scale (1=never, 5=always).35 36 The BHLS has been shown to correlate with the Short Test of Functional Health Literacy in Adults and the Rapid Estimate of Adult Literacy in Medicine as criterion standards.35–37 Following acceptable thresholds, the items were summed and dichotomised into low (total score ≤21) or medium–high (total score >21) health literacy.36 38
Minority status was based on self-identification. Minority patients in this study were members of the Arabic-speaking population (n=211) or self-identified native Russian speakers (immigrants from the former Soviet Union who immigrated to Israel mainly from the 1990s onward; n=198). Non-minorities were those belonging to the general (Hebrew speaking) Israeli populations (n=190). Based on patients’ self-identified native language, we coded Hebrew speakers as 0 and minorities (Arabic or Russian speakers) as 1.
Hospital discharge practices (parallel mediators)
Teachback
We used the patient-reported incidence of teach back, ‘Did a doctor or nurse ask you to repeat their instruction in your own words?’39 rated on a five-point Likert-type scale (1=not at all, 5=very much).
Providers’ culture competence
We used a patient reported measure of perceived Provider Cultural Competency40 that includes judgments of cultural knowledge, awareness and skills of providers,41 adapted to include six questions, answered using seven-point scales (1=not at all, 7=very well) (Cronbach’s alpha=0.93).
At-discharge provisions of care
As previously reported,22 indication of language concordance between the patient and the discharge physician or nurse (0=no, 1=yes) and on caregiver presence at discharge (0=no, 1=yes) was used to construct a four-level categorical variable. The provisions of care variable was coded as 1=none, 2=language concordance only, 3=caregiver present only and 4=both language concordance and caregiver presence.
Care-transition preparedness (serial mediator)
The care transition measure (CTM) was used to assess patients’ care-transition preparedness.42 The CTM was previously translated into Hebrew, Arabic43 and Russian and adapted for use in the Israeli healthcare system context by removing three items to generate the CTM-12 (which has shown good internal consistency; Cronbach’s alpha=0.93).16 22 Higher scores on the 0–100 CTM scale represent better care-transition preparedness.
Control variables
Health status was assessed using the SF-12 V.2,44 as reported by the patients. Higher scores on the 0–100 (Physical Component Score (PCS) and Mental Component Score (MCS)) subscales represent better physical/mental health.
Daily functioning was assessed using the Katz Index of Independence in Activities of Daily Living (ADL) scale45, as reported by the patients (0=independent, 12=dependent).
We also obtained information from Clalit’s data warehouse on age, length of stay, number of chronic conditions and hospitalisations during the year before the index hospitalisation.
Statistical analysis
To explore the mediation mechanisms and estimate the magnitude and significance of the direct and indirect effects, a parallel and serial multiple mediator model was specified (pp 180–182, PROCESS, Model 80)46 using the multiple X procedure (chapter 4, pp 144–145).46 As shown in figure 1, the model tests the pathways through which multiple predictors (cultural factors, Xi) simultaneously affect the parallel mediators (hospital discharge practices, M1–M3) which, in turn affect the serial mediator (care-transition preparedness, M4),47 which together affect the outcome (readmissions, Y), controlling for known covariates at the outcome level48: ADL, PCS, MCS, age, length of stay, prior hospitalisation and number of chronic conditions. The direct effects of cultural factors (Xi) on readmission (Y) when all mediators (M1–M4) are included are depicted by C′i; the indirect effects of Xi on Y through the mediators (M1–M4) are presented by the product of each path (ai X di X bi) (figure 1).
Parallel and serial mediation model* (based on Hayes’ Model 80, p 180).46 Predictors (Xi): cultural factors (X1: internal health locus of control (HLC); X2: powerful others HLC; X3: chance HLC; X4: family collectivism; X5: health literacy; X6: minority status); parallel mediators (M1–M3): hospital discharge practices (M1: teach-back method; M2: providers’ cultural competence; M3: at-discharge provisions (care giver presence and/or patient–provider language concordance); serial mediator (M4): care-transition preparedness (assessed by the care transition measure (CTM)). Outcome: 30-day readmission. Association between each predictor and each of the parallel mediators: ai1–ai3. Association between each predictor and the serial mediator: ai4. Association between each of the parallel mediators and the outcome: b1–b3. Association between the serial mediator and the outcome: b4. Association between each of the parallel mediators and the serial mediator: d1–d3. C′i: direct effects of cultural factors (Xi) on readmission (Y) when mediators (M1–M4) are included in the analysis; ai X di X bi = indirect effects of Xi on Y through the mediators (M1–M4) (product of multiplying coefficients of each path); ∑ai X di X bi=total indirect effects (sum of all the products of the specific paths for each predictor). *Control variables (clinical, personal and health service use characteristics) are not shown.
The analysis was performed as following: initially, descriptive statistics were calculated for the entire sample and for readmitted and non-readmitted groups. Multiple logistic regressions were used to measure the association between each predictor and the outcome, controlling for all covariates (total effect). Simple logistic regressions were calculated between all other variables and the outcome. The association between each predictor and each of the mediators (M1–M4) and between each of the parallel mediators (M1–M3) and the serial mediator (M4) were explored using simple Pearson correlations. Independent t-test analyses were used to determine the differences between the binary predictors’ categories for each continuous variable and χ2 were used for categorical data.
To test for the significance of the indirect effect, bootstrap CIs using the percentile method (based on 10 000 resamples) were used with a 0.05 criterion for rejection (ie, 95% CI considered statistically significant when the percentile bootstrap 95% CIs do not include zero).46 To compare among several parallel pathways, which are found to be specifically indirectly associated with the outcome, the assessment of the relative strength of each pathway (eg, X1→ M1 → M4 →Y vs X1→ M2→M4→ Y) was performed via contrast analysis.46 Significant contrasts are considered statistically significant when the percentile bootstrap 95% CIs do not include zero, indicating whether one of the parallel mediators has a significantly stronger indirect effect than the other mediators in its association to the outcome (readmission).
The discriminatory power of the model between readmitted patients and those not readmitted was calculated using the C-statistic with 95% CIs.49 Analyses were performed using IBM SPSS Statistical package V.23.0 and the PROCESS macro (V.3.1, model 80).46
Results
Descriptive statistics and intercorrelations
A total of 1013 hospitalised patients at six internal-medicine wards were invited to participate in the study. As previously reported,22 23 of the 675 baseline participants, 338 (33.3%) refused participation because of fatigue or not feeling well enough to complete a questionnaire, and 76 were lost to follow-up because of: death during the index hospitalisation (n=14), transfer to a different unit or different hospital (n=21), length of stay of more than 30 days (n=2) or refusal to participate in the follow-up telephone survey (n=39). Those lost to follow-up were not different from the final study sample (599 patients) in their demographic characteristics (age, sex and native language).
The table 1 presents characteristics of participants for the entire sample and for those with and without readmission. One hundred and one participants (17%) were readmitted within 30 days of discharge. Participants who were readmitted rated their internal HLC and powerful others HLC significantly lower, and their chance HLC significantly higher, and had lower levels of family collectivism and lower health literacy than non-readmitted patients. At the discharge briefings, there was less use of the teach-back method, providers’ cultural competence was rated lower and there were fewer at-discharge provisions in readmitted versus non-readmitted patients. Readmitted patients also had lower CTM scores than non-readmitted patients (p<0.01). Additionally, readmitted patients had one more hospitalisation in the previous year, had more chronic conditions and were of poorer physical, mental and functional status than non-readmitted patients (p<0.01).
Sample characteristics for entire sample and by readmission status
As shown in table 2, all cultural and hospital discharge practices were significantly and positively associated with the CTM (serial mediator). Almost all the cultural factors showed significant associations with the parallel mediators (M1–M3), with the exception of: internal HLC and powerful others HLC with teach back (M1) and at-discharge provisions (M3); family collectivism with the teach back (M1) and providers’ cultural competence (M2) and chance HLC and at-discharge provisions (M3) (table 2). The table 2 also shows significant differences in means of all study variables between low and medium–high health literacy levels except for powerful others HLC and teach back; and between minority and non-minority patients, except for family collectivism. Additionally, minorities were more likely to have low health literacy (χ2=62.31, p<0.001).
Bivariate analysis of predictors and mediating variables
Multiple mediation model
As shown in table 1, the multiple logistic regressions for each predictor (controlling for covariates) indicated that all cultural factors (internal HLC, powerful others HLC, chance HLC, family collectivism and health literacy), with the exception of minority status (B=0.002, p>0.05), were associated with 30-day readmissions (total effects, p<0.05).
The multiple mediation analyses are presented in figure 2. The unstandardised regression coefficients and their significance are indicated along each line. As shown in figure 2, all cultural factors are associated with the at-discharge provision of care, with the exception of internal HLC. Minority status (vs non-minority) is associated with all three hospital discharge practices (teach back, providers’ cultural competency and the at-discharge provisions). Low health literacy (vs medium-high) is associated with providers’ cultural competency and at-discharge provisions of care. Additionally, all cultural factors and all hospital discharge practices (parallel mediators) are associated with care-transition preparedness (serial mediator). The coefficients for each association differs in magnitude (eg, minority→ at-discharge provisions: B=−1.35 and minority→CTM: B=12.15) due to differences in the range of the scale values of each of the variables. Non-significant coefficients are shown in the online supplementary table S1.
Supplemental material
Results for the parallel and serial multiple indirect effects model of the cultural factors on readmission. The model shows significant unstandardised regression coefficients (non-significant paths are shown in online supplementary table S1). Dashed lines highlight non-significant relationships; R2: r squared for multiple linear regression; Nagelkerke’s R squared indicated the power of explanation of the logistic model. All the scales of the predictors and mediators are continuous except for (health literacy (1: medium–high, 0: low); minority (1: yes, 0: no)); At-discharge provisions included: patient–provider language concordance and/or caregiver presence at discharge briefings; the estimated indirect effects are presented in the online online supplementary table S2; *p<0.05. ADL, activities of daily living; CTM, care transition measure; HLC, health locus of control; MCS, Mental Component Score; PCS, Physical Component Score.
The multiple mediation analysis indicated that when the mediators (hospital discharge practices and care-transition preparedness) were included, the direct effects (C′i) of the cultural factors on readmission were no longer significant (see online supplementary table S2). In the final model, which included all mediators, minority status (OR=0.39; B=−0.95; p=0.02), CTM score (OR=0.96; B=−0.04; p=0.003) and the control variables: hospitalisation during past year (OR=1.36; B=0.30; p<0.001) and length of stay (OR=0.90; B=−0.11; p=0.016) were directly associated with 30-day readmission.
The percentile bootstrap estimates of the total and specific indirect effects were estimated and are also presented in the online supplementary table S2. All the specific indirect effects of each significant pathway in the model (as shown in figure 2) were statistically significant (this can be interpreted as significant mediation as the 95% CIs do not contain zero), confirming that the proposed constructs: hospital discharge practices and care-transition preparedness mediate the association between all cultural factors and 30-day readmission. Moreover, as noted above, the total effect of minority status was uniquely not significantly associated with readmission (B=0.002, p>0.05). Examination of the indirect effect of minority status on readmission (online supplementary table S2) shows that the effect on readmission was significantly driven by all the mediators (M1–M4). When we included the mediators, we found a significant strong direct and inconsistent indirect effects between minority status and readmission (B coefficient=−0.95; p=0.021). Because high-quality discharge processes were associated with higher care-transition preparedness (and, in turn, reduced readmissions), this reflects a situation46 (pp 117–118) in which the relationship between an independent variable (ie, minority status) and a dependent variable (ie, 30-day readmission) is strengthened when other variables are included (ie, mediators).47 50
Results of all pathways (online supplementary table S2) show that care-transition preparedness is a strong significant serial mediator between all cultural factors, the hospital discharge practices and 30-day readmission. This shows that given any type of cultural factor examined in this study, tailoring the hospital discharge practices to patients’ unique cultural characteristics contributed to patients’ overall assessment of their care-transition preparedness, which in turn, contributed to readmission reduction.
Finally, the absolute contrast analyses of the significant specific indirect effects of the parallel mediators on readmission through the CTM shows that the strongest mediator for the cultural factors health literacy and minority status was the at-discharge provisions of care (online supplementary table S2: significant indirect effect of paths 6–7 in the health literacy analysis and significant indirect effect of paths 5–7 and 6–7 in the minority status analysis). This implies that when language concordance and caregiver presence were available, this significantly contributed to reduced readmissions for patients with low health literacy and minorities.
The model accounted for 68% of the variance (ordinary least squares R2) in CTM assessment (by the cultural factors and the parallel mediators (M1–M3)). Additionally, the Nagelkerke’s R-squared of the prediction of 30-day readmission (by the cultural factors, mediators (M1–M4) and the control variables) was 30%. The C-statistic for the final multilogistic regression model was 0.82 (95% CI 0.78 to 0.86), indicating good discriminatory power.
Discussion
This prospective study shows, for the first time, that the association between patients’ cultural factors and 30-day readmission is mediated by hospital discharge practices and by care-transition preparedness, while accounting for a wide range of well-known clinical and administrative risk factors. It is now well established that high-quality patient discharge involves the provision of instructions for ongoing care and self-management, provided in a way that is patient centric and tailored to meet patients’ needs.8 17 51 Nonetheless, studies which assess patient perceptions are often focused on the dominant patient population, excluding linguistic minority patients who cannot answer questionnaires in the dominant language.12 Our results add to current knowledge by revealing the pathways by which high-quality discharge processes, tailored to culturally and linguistically diverse groups, mitigate the effects of culture on readmission.
Our findings also showed that while the total effect of minorities on readmission was not significant, when minorities received culturally competent care, teach back was used, there was language concordance with their providers and caregivers were present, and this resulted in significant higher ratings of their care-transition preparedness and overall reduced readmission rates. This suggests that for minority patients, compared with the general population, it is especially important to provide a patient-centric culturally and linguistically tailored care-transition process in order to prevent readmission.
Our study adopted a comprehensive approach to the assessment of cultural factors based on the understanding that culturally appropriate care is grounded in the cultural makeup of populations and that these relate to health behaviours.52 We show that when patients are characterised by a high external HLC associated with powerful others or chance and by high family collectivism, their discharge briefing is more likely to include the presence of caregivers and language-concordant care, and that this, in turn, contributes to care-transition preparedness and to reduced readmissions. Conversely, the relationship between internal HLC and readmissions was mediated only by perceived care-transition preparedness. These findings are supported by the understanding that when patients have high perceived internal control, their perceived abilities for health improvement increase and lead to fewer readmissions; when patients are characterised by high external locus of control, they require more facilitation of the discharge process to achieve similar favourable outcomes.53
Our findings that health literacy was inversely associated with the at-discharge provisions of care and providers’ cultural competency indicated that at lower health literacy levels, care was culturally and linguistically tailored. Strategies aimed at improving the discharge process and reducing readmissions for patients with low health literacy, such as the use of teach back, of language which is free from medical jargon, and of personally tailored messages, have been demonstrated to improve patients’ understanding of their disease and its treatment and have improved patients’ postdischarge care practices.54
Finally, our mediation analyses demonstrate that the care-transition preparedness was a strong serial mediator, beyond the contribution of the hospital discharge practices included in this study. These findings contribute to understanding of the pathways by which care-transition preparedness is associated with reduced readmissions. Previous studies have shown equivocal findings on the relationship between the care-transition preparedness and readmissions.8 9 Yet, these studies did not account for the wide range of cultural factors included in the current study, nor did they examine the pathways by which discharge practices affect care-transition preparedness.
Strengths and limitations
This study has several limitations. First, to ultimately ascertain the effect of discharge practices on readmissions, in patients from diverse cultural backgrounds, an experimental design should be employed. Nonetheless, previous research has shown that a patient-centric discharge process is associated with better overall preparedness of the patient for his/her postdischarge care especially in minority populations.15 Second, we collected information on the hospital discharge practices and on the care-transition preparedness at the same time point (the telephone follow-up). Nonetheless, the hospital discharge practices were assessed via patients’ reports on practices occurring at the day of discharge (eg, teach back was performed, caregiver was present) and the care-transition preparedness assesses the potential results of such processes (eg, ‘When I left the hospital, I clearly understood how to manage my health’). Additionally, the Israeli society and healthcare system may not be representative of other countries and other healthcare systems. For example, on discharge, each patient, in any Israeli hospital, receives a discharge letter detailing their postdischarge care plan. Yet, international research exemplifies the wide-ranging applicability of findings on care transitions of patients from diverse minority groups, such as those presented here.17 21 Finally, our outcome variable classifies any readmission as a negative outcome, demonstrating a 17% readmission rate, somewhat lower than the reported 19% average for the adult Israeli population,55 yet reflecting the lower overall readmission rates of Clalit hospitals. Others have called for measuring avoidable readmissions as some readmissions cannot be prevented.56 Nonetheless, as there is no consensus on which readmissions are in fact avoidable,57 future studies should aim for a more uniform definition, especially in multicultural diverse populations.
In conclusion, providing high-quality, patient-centric transitional care entails delivering care in a manner tailored to patients’ cultural and linguistic needs. This study, for the first time, links health-system processes and cultural/linguistic patient characteristics and examines their combined effect on hospitalisation outcomes. An understanding of the entire range of factors and their effects on the quality of care transitions may guide policy and practice in improving the quality of care by tailoring the discharge process to patients’ personal and cultural needs. Future studies should test interventions aimed at improving transitions of minority patients through tailored discharge-planning processes and examining the impact on the care-transition experience and on readmissions.
References
Footnotes
Contributors All authors meet the criteria for authorship and all those entitled to authorship are listed as authors. NRG drafted the manuscript, analysed and interpreted the data and obtained funding for the original study. ES participated in drafting the manuscript and interpreting the data. BT and RB provided critical revisions for the manuscript. All authors approved the final version of the manuscript.
Funding This study was funded by Scholarship from the Israeli Council for Higher Education (Planning & Budgeting Committee) for excellent Arab and minority doctoral students.
Competing interests None declared.
Patient consent for publication Not required.
Ethics approval Institutional review board approvals were received from the medical centre, from Clalit’s central ethics committees and from the University's institutional review board.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement No data are available.