What are the best practices for placement of SQL when checking for duplicates? WHERE clause JOINS, beginning or end?

Where do you put a checking for and removing duplicates query when you need to remove all extras. The goal is to check name, date of birth, address of person.  Then to make sure that I am getting only the last person from their most recent visit or date time stamp occurrence. (We'll call it their most recent date of service) or in this case "s._serviced_date"
I think it will look something like this
Select Name, Address, Date of Birth From Personnel
Where Max(Service_Date) ----Do I need any date formatting here or just the field?
Group By Name, Address, Date of Birth
Having Count(*) >1
Please feel free to check that. Especially If MAX() has to have special date format. May not need any?

THE BIG QUESTION HERE IS WHERE DO I PLACE IT AND WHY. WHAT ARE THE BEST PRACTICES OR WHAT IS GOOD POLICY TO DO IT ONE WAY OVER ANOTHER? the following is a mockup of the query. 

declare @from_dob date = dateadd(year,-45,'20210101')
declare @to_dob date = dateadd(year,-40,'20211231')
declare @research_date date = '20150101'
declare @region varchar(100)
        select @region = dbo.z_bi_get_region(db_name())
 
select 
@region as region
, personnel_id
, last_name
, first_name
, gender_code
, birth_date
,datediff(year,birth_date,cast(getdate() as date)) as age_at_run_date
, mobile_phone_number
, home_phone_number
, primary_email_address
, primary_address_line1
, primary_address_line2
, primary_address_city
, primary_address_state
, primary_address_zipcode
, death_flag 
 ,isnull(s1.disease_found,'N') as disease_found
 ,@from_dob as from_DOB, @to_dob as to_DOB, @research_date as research_date
 
from
   c_personnel pers
   outer apply
   (
     select top 1 'Y' as disease_found from c_research s
       inner join c_research_item si on s.research_key=si.research_key and si.active_flag='Y'
       inner join l_procedure_code pc on si.procedure_code=pc.procedure_code 
       inner join l_disease_type dt on pc.disease_type_code=mt.disease_type_code and dt.disease_type_code in ('Cancer','Heart','Kidney',’Lung’,'Accident',’Neurological’)
       inner join l_location st on s._location_code=st.location_code
     where 
       pers.personnel_key=s.personnel_key
       and s._serviced_date  >= @research_date
       AND st.location_code NOT IN ('HP','CLNK') AND st.practice_code <> 'DocNABox'
   ) as s1
where 
   pers.birth_date >= @from_dob 
   and pers.birth_date < dateadd(d,1,@to_dob)
   and pers.master_personnel_key is null
   AND pers.issuer_of_personnel_id = 'system' 
   and pers.gender_code='M'
   and isnull(pers.death_flag,'N')='N'
   and isnull(s1.disease_found,'N')='N'
  order by datediff(year,birth_date,cast(getdate() as date))

THE BIG QUESTION HERE IS WHERE DO I PLACE IT AND WHY. WHAT ARE THE BEST PRACTICES OR WHAT IS GOOD POLICY TO DO IT ONE WAY OVER ANOTHER? the following is a mockup of the query.