Three Strategies for Implementing Motivational Interviewing on Medical Inpatient Units

General medical hospitals provide care for a disproportionate share of patients who abuse or are dependent upon substances. This group is among the most costly to treat and has the poorest medical and substance use outcomes. Motivational interviewing(MI) is a well-recognized, evidenced-based substance use treatment that has been adapted for use as a brief intervention in health care settings. MI is applicable to many health-related behavioral problems, and can be taught to a broad range of health care clinicians. However, it is unclear which implementation strategies will lead to the efficient and proficient uptake of MI in general medical settings, such as medical inpatient units.

Primary care clinicians have multiple practice demands and time constraints. New practices have the greatest chance of being implemented if they are simple and compatible with existing workflows and systems. Two widely used strategies to bring specialized practices into use within general hospital settings are the "see one, do one" apprenticeship model of training and use of consultation-liaison (CL) services. "See one, do one" has been a modus operandi in medical education for centuries and relies upon a competency-based supervision training approach. While it has been empirically validated in the specialty addiction field, less controlled testing of this implementation strategy is available in general medical settings. The apprenticeship approach requires that appropriate patients and trainers are available with high flexibility for teaching and supervision; when applied to behavioral counseling approaches, this may be seen as incompatible with the medical role and time constraints of clinicians. In contrast, ordering MI through CL is a relatively simple, minimally burdensome process and highly compatible with the way clinicians secure other specialist services for their patients in the hospital.

We propose to conduct a randomized controlled implementation trial using mixed quantitative and qualitative methods to examine the effectiveness of three different strategies for integrating MI into the practice of healthcare providers working within Yale New Haven Hospital's internal medicine hospitalist service and other general medical inpatient units. Specifically, we will randomize 40 healthcare providers to one of three conditions: (1) a continuing medical education workshop that provides background and "shows" healthcare providers how to conduct MI (the control condition, called SEE ONE); (2) a "see one, do one" apprenticeship model involving workshop training plus live supervision of bedside practice (DO ONE); and (3) ordering MI from CL after learning about it in a workshop (ORDER ONE). Following the respective MI trainings, each healthcare provider will be assessed for the provision of MI to 40 study-eligible inpatients, recruited by the research team after admission to our general medical units.We hypothesize that the percentage of MI sessions delivered by providers to study-eligible inpatients would be higher in both Do One and Order One than See One. We also hypothesize that providers in both Do One and Order One would conduct MI sessions with greater integrity (i.e., adherence to core components of MI and delivery of them with competence) than those in See One.

Please note, as an implementation trial, the primary outcomes for this study focus on provider behaviors, namely, uptake of MI sessions with patients and the adherence and competence in which they conduct MI sessions. No outcome data will be collected and reported at the patient level.

For Healthcare provider participants:

Inclusion criteria

• Assignment to one of the general medical inpatient units during day-time shifts; intensive care units will be excluded given the morbidity of patients in this setting.
• Volunteer to serve as study clinicians, attend a workshop about MI, and possibly receive live supervision.
• Agree to all procedures of this trial (randomization to training condition and of assigned patients, audio recording MI sessions, and completing assessments).

Exclusion criteria

• Have been formally supervised to use MI with patients on the units.
• Intend to give notice that they plan to leave the hospital or are scheduled for medical or family leave such that they will not be able to interview 40 patients during the study period.

For patient participants:

Inclusion criteria

• Are 18 years of age or older.
• Acknowledge use of a substance within past 28 days and meets screening criteria consistent with substance (illicit drugs, licit drugs that are used in a non-medically indicated fashion, alcohol, or nicotine) use disorder.
• Are willing to consent to audio recording of interview with the provider or CL clinician.

Exclusion criteria

• Have an altered mental status such as delirium, encephalopathy, dementia or mental retardation or a score on the Confusion Assessment Method > 0 since this would impair provision of consent and ability to participate
• Inability to speak English. Most of providers are mono-lingual English speakers, and all MI integrity raters only speak English. We therefore do not have the capacity to include Spanish-only speaking patients in the study.
• Stroke (that precludes participation)
• Resides in a nursing home, skilled nursing facility or Hospice Care
• Receiving palliative care
• Deaf
• Unable to speak lucidly
• Previous participation in the protocol

An information sheet was requested and approved for a subset of patient subjects. This is due to the study being conducted within an acute medical inpatient unit, where conditions that might limit a person's ability to sign the consent form may occasionally occur. This subset of patients includes: patients that are physically unable to write (i.e. hand tremors, spinal cord injury, stroke that precludes signing, broken hand, broken shoulder, muscular dystrophy and other physical ailments preventing a patient from physically signing), unable to see (i.e. legally blind, uncontrolled type 2 diabetes mellitus which led to blurred vision), unable to read (i.e. patient does not have their glasses on them).