This is begging the question, but we still use it because people are on it and it works. It's binding affinity to the MOR is higher than most other opioids (explaining blocking euphoric effects of other opioids for the same reason buprenorphine does), and in the 1960s people noticed that it would quell cravings without producing changes in consciousness or levels of functioning (people could still work, appropriately interact, etc).
Compare with Buprenorphine methadone has better treatment retention rates but access remains a problem (meaning, you need a licensed clinic). Additionally dosing and prescribing methadone takes more training and understanding of pharmacology (given it's variable but long half life, it's storage in peripheral tissues with delayed release in blood stream, significant interactions with other drugs, potential for cardiotoxicity etc).
A lot of older OBGYNs prefer to prescribe it during pregnancy due to familiarity and the bulk of literature advocating for its use during pregnancy because this is what they used during the heroin epidemic in the 1970s. However this too is changing given that Buprenorphine is safe, easy to use, as effective (though again greater retention in treatment), and more straightforward in pregnant patients. Additionally it produces a less severe neonatal abstinence syndrome.
One of the reasons I like Buprenorphine a little better is because, as one of my attendings says, the two things that are guaranteed with opioid addiction are constipation and depression. Work from Gold and Kleber in the 1980s shows that chronic MTD (and by extension other opioid agonist) treatment suppresses release of endogenous opioids, and emerging optogenetics work is elucidating the link between the opioid system and dysphoric/anehdonic states. The KOR antagonism of Buprenorphine can really help patients with this, but again, pharmacotherapy is only a piece of the puzzle.