Here's a chance:
Create a table and separate the ids and possible treatments with separate records.
CREATE TABLE 'table_agendamentos' (
'id_table_agendamentos' int(11) NOT NULL,
'id_agendamento' int(11) NOT NULL,
'id_tratamentos' int(11) NOT NULL,
PRIMARY KEY ('id_table_agendamentos')
);
INSERT INTO 'table_agendamentos' VALUES
(1,1,1),(2,2,1),(3,2,2),(4,2,3),(5,3,2),(6,3,3);
Keep the table below:
CREATE TABLE 'tabe_tratamentos' (
'id_tabe_tratamentos' int(11) NOT NULL AUTO_INCREMENT,
'nome' varchar(45) NOT NULL,
PRIMARY KEY ('id_tabe_tratamentos')
);
INSERT INTO 'tabe_tratamentos' VALUES (1,'Facial'),(2,'Corporal'),(3,'Outros');
Now to return what you want to do the following SQL:
select * from tabe_tratamentos where id_tabe_tratamentos IN (
SELECT id_tratamentos FROM table_agendamentos where id_agendamento = 2)
or
Select TT.nome from tabe_tratamentos as TT
Inner join table_agendamentos as TA ON TA.id_tratamentos = TT.id_tabe_tratamentos
where TA.id_agendamento = 2;
All results:
Select TA.id_agendamento, GROUP_CONCAT(TT.nome) from tabe_tratamentos as TT
Inner join table_agendamentos as TA
ON TA.id_tratamentos = TT.id_tabe_tratamentos
GROUP BY TA.id_agendamento;