Reducing Sedatives in Hospital Study

Inappropriate prescription of unnecessary and potentially harmful medications continues to be a widespread problem for hospitalized patients. We aim to study the prescription patterns of several classes of medications featured both in the Canadian Choosing Wisely campaign and the Beers Criteria for medication1,2. Specifically, we will be looking at the patterns of sedatives (includes benzodiazepines, sedative-hypnotics, and sedating antipsychotics) prescriptions over time.

Use of sedatives in community or hospital settings has been shown to range from 50-80%. Extensive study has linked their use and falls, likely due to their negative effect on postural control and cognition3. The odds ratio for falls in elderly patients taking benzodiazepines has been estimated at 1.42 (95% Confidence Interval, 1.20-1.71) and even higher patients over the age of 65 for falls in patients taking a benzodiazepine 4. This association also holds true for hospitalized patients, particularly if first prescribed in hospital5. Similarly, antipsychotics have been shown to increase the risk of CVA (cardiovascular accident) and mortality in elderly patients and are not recommended as first-line therapy for insomnia.1 In addition, sedatives are associated with increased mortality and morbidity such as hip fractures. Sedative prescriptions that originate in hospital are perpetuated after discharge with over 3% of sedative-naïve patients leaving hospital with a new prescription. Half of these patients go on to become chronic users.6 Guidelines now strongly recommend against the use of sedative medication as first-line therapy for insomnia. There is evidence to support a non-pharmacological approach to iatrogenic insomnia in hospital that involves environmental cueing (noise and light reduction, warm beverages at bedtime) and minimizing interruptions.7 We aim to determine how closely these recommendations are adhered to on medical-surgical wards at five local hospitals (Sinai Health System, Toronto General Hospital, Toronto Western Hospital, Sunnybrook Health Sciences Center, and St. Michael's Hospital); and to observe trends in prescribing before and after hospital-wide quality improvement initiatives.

Rationale for this study

Despite recommendations against the regular use of sedatives, these drugs continue to be widely prescribed in the inpatient setting. This study will provide baseline information on the pattern of sedative use on medical-surgical wards. It will also identify changes in prescribing patterns following institutional quality improvement initiatives. Participating hospitals will implement a sedative reduction sleep bundle including the following best practices:

• nursing, physician and patient education such as information sessions, hand-outs, posters, and email alerts
• implementation of sleep-friendly environmental changes (e.g., minimize noise and disruptions overnight)
• monitoring of the impact of environmental changes on patient's sleep through the use of clinical sleep surveys (administered by front-line staff)
• clinical pharmacists engagement to assist clinical teams in carrying out best-practice with regards to appropriate sedative prescribing
• revisions of existing order-sets to reflect the appropriate ordering

Potential risks and benefits to participants

There is no potential for physical harm to participants in this study. There is sufficient scientific evidence that guides clinicians away from prescribing sedatives and antipsychotic drugs for insomnia in the hospital setting. In fact, there is potential benefit to welfare in this quality improvement initiative. This is a quality improvement initiative that seeks to implement best practices to reduce avoidable harm to patients. Ethics review is requested at each participating site for access to health records and for the dissemination of results demonstrating the effectiveness of the sedative reduction sleep bundle on reducing sedatives without diminishing sleep quality in hospital.

The risk of release of personal health information is minimal as the data will be de-identified and all standard security measures will be applied. The principal investigator of each site will assign patients with a study number and remove identifying data (exceptions: age and sex) before exporting de-identified data from host hospital. Site PIs will maintain the study identifier key in locked offices on hospital-secure servers. Only the study principal investigator will have access to de-identified patient-level data, which will be reported as aggregate data. Data sharing agreements will be completed. Any potential benefit would come in the form of potential improvements to the health care system and safer future prescribing practices.

Patient population to be studied

The study population will be any patient admitted to medical-surgical units.