Patient-oriented education and medication management intervention for people with decompensated cirrhosis: study protocol for a randomized controlled trial

The primary aim of the study is to determine whether the intervention (medication review, reconciliation, and additional education by a clinical pharmacist) reduces the frequency (counts) and severity (measure of clinical significance) of discrepancies between patient-reported and medical record-documented “current” medications. The secondary aims are to describe the impact of the intervention on medication adherence, quality of life (QoL), medication and illness beliefs, patient knowledge of disease and self-management tasks related to decompensation history and prescribed therapy, and the frequency of hospitalization for management of cirrhosis-related complications, non-cirrhosis-related complications, and survival. The sustainability of the intervention in the current health care system in relation to costs and cost-effectiveness will also be evaluated.

The study has a randomized, controlled, parallel-group design and will be conducted at a single site, the Hepatology Outpatient Centre at the Princess Alexandra Hospital (PAH), located in Brisbane, Australia. The Gastroenterology and Hepatology Department at the PAH is one of the largest hepatology centers in Australia and the only liver transplant center in Queensland. The hospital’s local catchment area is 3,856 km². Additionally, many patients travel from regional and remote areas, including interstate to access specialist services. Consequently, the Hepatology Outpatient Centre is responsible for the ambulatory care of a significant proportion of southeastern Queensland patients with chronic liver disease (CLD).

The Hepatology Outpatient Centre’s appointment database will be prescreened for eligible participants who are scheduled for routine review by a hepatologist. Persons identified as eligible will be contacted via telephone in the week preceding their appointment to discuss the study and offer participation. Participants will be given instructions to arrive 15 minutes early for their appointment, bring all medications to the clinic, and ask their carer or responsible family member to accompany them if appropriate. Consented participants will be randomly allocated to one of the two study arms: (1) intervention or (2) usual care.

An independent person at the PAH will prepare individual envelopes labeled with the randomization number and containing the corresponding study arm. Allocation concealment will be achieved by use of sequentially numbered, sealed, opaque envelopes that will be stored and opened in order by a third party located within PAH.

Figure 1 describes the planned study flow, including data collection time points. A Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) checklist of standard protocol items [18] addressed within the study protocol is provided in Additional file 1.

All participants will complete a baseline survey prior to their scheduled outpatient clinic review. Validated questionnaire tools will be used to determine baseline medication adherence, medication beliefs, illness perceptions, and QoL, in addition to patient knowledge of important dietary and lifestyle modifications. The survey will be completed with a study coordinator over the telephone or in person in a private clinic room, or it will be self-completed by the participant in the clinic waiting room. Patient carers or family members will be involved if necessary.

Usual care participants will subsequently receive routine review and education by their hepatologist according to the current model of care in the hepatology clinic. Intervention participants will be reviewed by a clinical pharmacist to receive the education and medication management intervention (described below) in addition to routine review and education by their hepatologist. Intervention participants will also receive three further educational contacts from the clinical pharmacist in addition to scheduled routine hepatologist review within the following 6-month period.

All participants will be resurveyed at follow-up (t ₃, approximately 6 months post-enrollment) to explore changes in their medication adherence, medication beliefs, illness perceptions, QoL, and knowledge. Patient outcome data will be collected at 52 weeks (see “Study endpoints and data collection” section below).

A multifaceted, clinical pharmacist-driven education and medication management intervention targeted to patients with decompensated cirrhosis was developed by a panel of clinicians (including a hepatologist, general physician/pharmacologist, clinical nurse, and clinical pharmacist). Education and monitoring points (Table 2) were agreed upon on the basis of clinical hepatology guidelines to augment and support information provided by hepatologists and other clinicians involved in patient care.

During the initial encounter with each intervention patient (and carer/family member if present), the clinical pharmacist will obtain a complete, reconciled list of current medications and identify MRPs. The pharmacist will use open and closed questions to ascertain adherence, other medication-taking behaviors, and medication needs on an individualized basis. Participating patients will receive medication- and lifestyle-related education and advice tailored to their needs (decompensation history and medications) and receptiveness (Table 2). The pharmacist will also collaborate with the patient’s general practitioner and hepatologist to optimize therapy, facilitate resolution of MRPs, and monitor issues where required.

Subsequent contact with participants will be made at 4–6 weeks (t ₁) and 12–14 weeks (t ₂) after the baseline interview to discuss changes in prescribed therapy; identify new MRPs and previously identified MRPs that are ongoing; reiterate advice provided at baseline; and provide additional relevant education about medications, lifestyle modifications, and self-management of cirrhosis to build on previous discussions. Promotion of a patient-clinician partnership will be central to each discussion to encourage functional health literacy and self-efficacy with regard to patient confidence to raise questions and discuss issues with health care providers.

During patient interviews, the clinical pharmacist will identify actual and potential MRPs and document these in the patients’ medical notes and on a structured data collection form. The pharmacist will then facilitate the resolution or monitoring of individual MRPs with other relevant clinicians as appropriate and document each outcome in patients’ medical notes and data collection forms.

The clinical significance of MRPs (and medication discrepancies) will be assessed by a clinician panel using a risk matrix (Fig. 2) modified from the Metro North Health Service Risk Management Framework [19]. The matrix assigns a composite risk of prospective severity and likelihood of potential harm, as well as duration of time until potential harm may occur, as a consequence of the MRP. The panel of clinicians will comprise at least one clinical pharmacist, hepatologist, and general physician/pharmacologist who have experience in managing patients with cirrhosis. MRPs will be de-identified, randomized, and independently assessed by at least two members of the panel. Relevant clinical information will accompany the de-identified MRPs, but not details of patient outcomes or MRP resolution, to ensure unbiased prospective assessment of potential harm. Consensus of individual rankings will be used to determine the final measure of potential harm. Where there is disagreement between individual rankings, a roundtable panel discussion will be called to facilitate consensus.

Intervention participants will receive education and advice tailored to their needs, receptiveness, and cognitive barriers. For example, a patient who has nonalcoholic steatohepatitis with well-controlled abdominal ascites and no other decompensation history would likely require a low-intensity intervention if he or she identified as adherent. Low-intensity intervention for this patient would consist of reiteration of important dietary modifications, monitoring of weight, and diuretic education. Conversely, a malnourished ex-alcoholic patient with a history of poorly controlled ascites, recurrent encephalopathy, and persistent hypotension will require high-intensity intervention. This would ideally involve inclusion of a carer/family member in the discussion to educate both parties on the potential precipitants and early signs of encephalopathy, self-titration of lactulose, and when to present to the emergency department. Pharmacotherapy education (including possible discussion of adherence aids), high-protein and low-sodium dietary modifications, daily weight monitoring (including when to contact the clinic if weight is changing significantly over short periods of time), and review of pharmacotherapy that may contribute to hypotension would also be conducted. Liaison with additional health providers may also be required.

After providing education and advice to intervention participants, the clinical pharmacist will use a structured data collection form to record topics that have been discussed, topics that were not discussed and reasons why, and whether a carer was present. Post hoc testing will be employed to explore study outcomes according to the intensity of the intervention received and the completeness of relevant information provided (see the “Data analysis” section below).

Using information from several sources (including the patient, general practitioner, pharmacy, own medications, carer), the clinical pharmacist will construct a reconciled list of current pharmacotherapy. The reconciled list of medications will be made available to other clinicians in the patient’s medical record, and a patient-friendly guide will be constructed using the approved statewide template. This guide will be printed and provided to the patient. The patient will be encouraged to take it to all health care appointments and use it at home to aid with compliance and memory. The guide will also contain additional medication management and lifestyle points related to prescribed therapy, such as low-salt diet reminders for patients who are prescribed diuretics in the management of ascites. Patients will be encouraged to manually edit the medication guide when therapy is altered.

Study outcomes, tools, and time points for measurement are outlined in Table 3. Follow-up will occur approximately 6 months post-enrollment (allowed range 140–224 days after baseline interview), according to outpatient hepatology scheduling availability. If a follow-up appointment is missed, a second appointment will be offered to the patient within the allowed time frame. If the second appointment is missed, the patient will be considered lost to follow-up for t ₃ survey endpoints.

A questionnaire will be used to collect data on the secondary outcomes of interest at baseline and follow-up (t ₃). Validated tools used to measure these endpoints will be included in the questionnaires with approval from their respective developers.

Study progress, collected data, and clinical cases will be reviewed periodically by the study investigators. Frequency and causation of participant hospitalization and mortality will be reviewed at 6, 12, and 18 months because the study group of interest is at high risk for these occurrences. If significant differences are found between the intervention and control groups, or if serious adverse events develop throughout the study, an independent data safety board will be convened to review the appropriateness of continuing the study in accordance with the Australian National Health and Medical Research Council National Statement on Ethical Conduct in Human Research [35].

Data will be coded and stored in a password-protected database and analyzed using a statistical software package (IBM SPSS Statistics version 20.0; IBM, Armonk, NY, USA). Study endpoints experienced for each treatment group will be formally analyzed on an intention-to-treat basis [36]. Important potential confounders will be included in a secondary multivariate analysis (e.g., hospital admission, mortality). Post hoc subgroup analysis will be used to further explore outcomes according to the intensity of the intervention and completeness of relevant personalized education received by intervention participants.

The study endpoints will be presented quantitatively (proportions, statistical differences between groups) and qualitatively (description of discrepancies, MRPs, and some patient-reported medication beliefs and illness perceptions). All tests will be two-tailed, and significance will be set at α = 0.05. Continuous and normally-distributed variables will be presented as mean ± SD. Differences between groups will be analyzed using one-way analysis of variance. Non-normally distributed data will be presented as median (range), and nonparametric tests will be used to analyze differences between groups. Categorical data will be analyzed using Pearson’s chi-square test or Fisher’s exact test. Pearson’s correlation or Spearman’s rank correlation coefficient will be used to identify linear relationships between normally distributed and non-normally distributed continuous variables, respectively.

Differences in the study endpoints of interest will be examined between the intervention and control groups at baseline and at follow-up. The endpoints will be examined at baseline to test for homogeneity between groups (effectiveness of randomization) and to identify clinical variables that may influence endpoints (e.g., Model for End-Stage Liver Disease score; Child-Turcotte-Pugh classification; decompensation history, including event date, type, severity, and recurrence; hospitalization frequency; age; sex; biochemical abnormalities; and other prognostic variables). Clinically significant differences at baseline between groups will be adjusted for during the final analysis. Changes in the frequency of medication discrepancies, self-reported adherence, QoL, medication beliefs, illness perceptions, and patient knowledge at follow-up will be examined to measure the effect of the intervention on study outcomes at group and individual levels. The primary endpoint of medication discrepancies will be explored using total frequency of instances and frequency of instances per patient and per medication group. Comparison between intervention and control groups at baseline and t ₃ will be made using total number of instances per group, total number of high/very high risk instances per group, mean/median number of instances per patient, and mean/median number of high/very high risk instances per patient.

Predictors of clinical outcomes will be identified using binary logistic regression to determine the OR (95% CI). A multivariable logistic regression model will be used to determine independent predictors for patient outcomes. The multivariable logistic regression model will be constructed using variables of clinical significance (severity of disease; decompensation history, including date, type, severity, and frequency of events; other prognostic variables) and statistical significance as determined by univariate analysis.

Time to first episode of decompensation and time to death (cause-specific and all-cause) will be assessed using Kaplan-Meier survival curves. The curves will be compared with the log-rank test statistic. Cox proportional hazards modeling will be used to calculate HRs, after testing the proportional hazards assumption. Patients will contribute person-time from the date of recruitment until the earliest of end of follow-up (12 months) or death (date of death to be obtained from medical records). Competing risks analysis will be performed in which we encode, as separate events, time to first readmission with a cirrhosis-related episode, death resulting from cirrhosis, or death resulting from other causes. This will be extended to multivariate analysis if there are a sufficient number of events.

It is anticipated that a proportion of participants (approximately 30%) will be lost from the study before completing follow-up assessments at t ₃. Although the sample size calculation has been adjusted to accommodate for this loss, subjects who have completed only one assessment will be included in the analyses to establish baseline parameters in both treatment groups. With respect to changes from baseline, data imputation will be considered. In decompensated cirrhosis, dropouts due to death are informatively censored, and thus the responses of patients with missing data will be modeled explicitly. Subjects will be matched using baseline demographic and patient-reported outcomes, repeating this technique for both intervention arms. Imputation methods such as a hot-deck imputation and iterative regression imputation will be used. The sensitivity of the results to each of these methods will be assessed to quantify the degree of bias present.

On the basis of a primary endpoint of a 50% reduction in medication discrepancies (3.12 ± 2.5 per patient [6] to 1.56 per patient), 41 participants will be required per treatment arm to achieve ≥80% statistical power. Median survival in decompensated cirrhosis is 2 years [37]. Therefore, a 6-month mortality rate between 10% and 15% among the study cohort is predicted. The Hepatology Outpatient Centre’s attendance failure/reschedule rate is approximately 20%. Therefore, a possible 30% loss will be accounted for, and at least 106 participants will be recruited.

In this study, we will evaluate the effectiveness of a single pharmacist in one general hepatology clinic in Australia, and applicability of findings to other sites will be dependent on similarities in patient demographics and models of care. An important component of this study is the provision of an intervention that is tailored to the patient’s needs, receptiveness, and cognitive barriers. As a consequence of the “reactive” nature of the intervention, the information provided will be selective, and delivery will be staggered over the course of the study. Perceived variability in delivery may be due to lack of clinical indication (i.e., not providing information on precipitants of hepatic encephalopathy to patients who have never had encephalopathy), time restraints, patient disinterest, or “information overload.” Where possible, advice and education points will be reiterated and built upon during future contacts.

There will be heavy reliance on patients’ self-reported information, such as the details of their local pharmacy and general practitioner for the process of medication reconciliation; self-reported adherence; QoL; beliefs; and perceptions. This and other clinical information will also be dependent on honest patient responses to the survey questionnaires. Although this is acknowledged as a limitation, patient reporting remains the most important method of information gathering in clinical practice in terms of eliciting history and clinical status. Furthermore, self-reported adherence using the MMAS-8 has been shown to be equally as accurate as and less onerous than other methods of adherence measurement, including pill counts. Where necessary, supplemental information will be sought from patients’ family members or other carers to corroborate patients’ responses if the study clinicians have concerns about patient safety in accordance with good clinical practice.

At the time of proof review, recruitment into the study is complete.