Homecare Agency-Randomized Trial of Web Implementation Strategy for Depression

MHTraining-Net uses a web-based platform and several long distance tools (e.g., e-learning modules, telephone or web consultation, telephone calls, toolkits, and discussion boards) to deliver four types of implementation activities: infrastructure development, training and education, quality improvement, and social networking. MHTraining-Net was developed to ensure that the Depression CAREPATH (CARE for PATients at Home), specifically, and evidence-based practices, generally, can be implemented widely and successfully across independent organizations within a sector of care.

PROCEDURES: The effectiveness of MHTraining-Net on the implementation of the Depression CAREPATH will be tested in an agency-randomized trial of 160 HHAs. All HHAs are clients of Brightree (formerly CareAnyware, Inc), a software company that offers web-based point-of-care clinical management and administrative support for HHA nationwide. HHA clinicians record relevant data that is transmitted directly to Brightree for clinical and administrative purposes, including quality improvement initiatives. As part of each HHA's contract with Brightree is permission for using these data for performance enhancement research.

Brightree has already programmed Cornell's Depression CAREPATH, a depression care management protocol, into its clinical management system so that all nurses use the protocol. To promote the use of the protocol, Brightree has posted two relevant videos (archived Patient Health Questionnaire-9 training web seminar training); demonstration of how to navigate the protocol within the software) on its user education website.

Brightree will randomize HHAs to either MHTraining-Net, (N=100) or to Enhanced Support (N=60). Enhanced Support will have access to only the two training videos until the end of the study when they will be invited to MHTraining-Net. These procedures represent Brightree's typical "roll-out" approach on introducing new training and products to their client HHAs.

The study will test its hypotheses using three sets of data:

1. Summary statistics describing the size and type of the 160 HHAs as well as their average nurse productivity over the 12 months preceding and following the study;
2. Routinely collected administrative data on patients age >65 who screen positive for depression over one year, including Medicare-mandated assessments, medication lists, transaction files, indicators of whether the depression care management protocol was used, and Medicare case-mix adjustment and reimbursement indicators. We are using to term 'patients' to refer to a full episode of patient care, defined from the Start-of-Care date to Discharge, recognizing that a small proportion of unique individuals may have more than one full episode of care. We will not be able to identify such individuals. These data will be aggregated at the patient level in preparation for performance enhancement analyses using Brightree's routine procedures. Brightree will code the data, remove identifiers, and transmit them to Weill Cornell on a regular basis using the Weill Cornell ITS Department secure servers and protocols.
3. Agency-level data on the use of the MHTraining-Net website (e.g., summary reports on 'hits' of each webpage) and agency-initiated contacts (telephone, email) through records maintained by Cornell project staff.