Johnston 2010.
| Study characteristics | ||
| Methods | RCT Unit of randomisation: hospital |
|
| Participants | Place of recruitment: 12 hospitals Numbers randomised: total: 3361 (I: 1464; C: 1897) % Completing final follow‐up: 80% Inclusion criteria: ischaemic stroke; Kaiser Permanente Medical Care Plan members with pharmacy benefits; age ≥ 40 years; acute hospitalisation for stroke Exclusion criteria: haemorrhagic stroke; discharged to hospice Type of stroke: ischaemic (100%) Mean age (SD): 72.9 (12.6) Gender (% women): 53% Ethnicity: non‐Hispanic white 66%; African American 14%; Asian/Pacific Islander 11%; Hispanic 7%; other/unknown 1% Socio‐economic or socio‐demographic status: members of Kaiser Permanente Medical Care Plan with "under‐representation of the very poor and wealthy" |
|
| Interventions | Intervention details (components, length, frequency): hospitals received support from a central coordinator in the development and implementation of standardised stroke discharge orders (discharge orders based on American Heart Association recurrent stroke prevention guidelines and included 1) statin prescription for all patients irrespective of cholesterol levels; 2) antihypertensive prescriptions for hypertensive patients; 3) warfarin prescription for patients with atrial fibrillation); 2 physician 'champions' (from neurology and hospital‐based medicine) from each hospital tailored discharge order and supervised implementation; 2 educational presentations delivered to healthcare providers (timing: development of discharge orders and 3 months post‐implementation) Location: Kaiser Permanente Medical Care Plan hospitals Mode of delivery: health provider education and pre‐printed stroke discharge orders Personnel responsible for delivery: central co‐ordinator and 2 physicians supervised implementation Timing post‐stroke: discharge from hospital Control: usual care without contact from study staff; some hospitals implemented their own discharge orders |
|
| Outcomes | 6 months: BP < 140/90 mmHg; combined cardiovascular risk factor control; adherence to secondary prevention medications | |
| General Information | Funding: Centres for Disease Control and Prevention, administered through the Association of American Medical Colleges Country of origin: USA Publication language: English |
|
| Notes | Analysis method: stated intention‐to‐treat Risk of bias: unclear |
|
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | "Participating hospitals were paired based on characteristics that could have impacted the success of the intervention, including patient demographics, hospital size, number of enrollees, and presence of a motivated stroke expert. Then, using a random number generator, 1 hospital in each pair was randomized to receive the intervention, whereas the other was randomized to usual care." |
| Allocation concealment (selection bias) | Low risk | "Participating hospitals were paired based on characteristics that could have impacted the success of the intervention, including patient demographics, hospital size, number of enrollees, and presence of a motivated stroke expert. Then, using a random number generator, 1 hospital in each pair was randomized to receive the intervention, whereas the other was randomized to usual care." |
| Incomplete outcome data (attrition bias) All outcomes | Low risk | Missing data reported by group Attrition: I: 1149/1464 (237 died; 78 lost to follow‐up); C: 1533/1897 (277 died; 87 lost to follow‐up) Judgement: reasons for missing data reported and review authors judge that they are unlikely to be related to study outcomes |
| Selective reporting (reporting bias) | Unclear risk | Protocol available and primary outcomes are reported in the pre‐specified way; some secondary outcomes not reported |
| Other bias | Low risk | The study appears to be free of other sources of bias |