Is it possible to capture the data from the Word document instead of the ECM forms?
It depends which application you are looking to use.
IDC only accepts image files, so you could convert the Word document to a PDF and process as usual.
ECM will accept almost any format, but it will specifically take Office documents with the option enabled in the ECM Client. In order for it to process the document as an image you will need to have Microsoft Office installed on the same machine where that ECM Client is running. Once in ECM you can use the OCR task to scrape the data from the Word document. NOTE: The OCR task is not dynamic at all so your Word document would need to be in the EXACT same layout every time to accurately scrape the data.
An alternative may be to create an XML map or CSV file to accompany the Word document so that you could use the other options within the ECM Client to pull the metadata into ECM without needing to use the OCR task.