Health and homelessness in Scotland: research

Annex C: Health Activity Datasets used in the Study

Accident and Emergency data (A&E2)

The A&E datamart was established in June 2007 to monitor the compliance of each NHS Board against the 4 hour wait standard. In July 2010, the A&E datamart was extended further to collect items such as diagnosis, several injury fields and an alcohol involved flag - which is used to identify whether the patient's alcohol consumption was a factor in the attendance. There are two types of data submitted to the A&E datamart: episode and aggregate level data. Sites that submit episode level data account for around 94% of all attendances at A&E. For more information, see the Accident and Emergency Datamart (A&E2) webpage.

Population Data
Data Controller:	NHS National Services Scotland
A record represents:	An A&E attendance
Data collection start:	June 2007. Rebuilt and re-launched to NHS Boards in January 2011
Frequency of collection:	Monthly
Number of annual records:	1,000,000+
Geography coverage:	Scotland
Population coverage:	All Emergency Departments and departments that provide Accident and Emergency (A&E) services in Scotland.

Study Data
Data period:	1 January 2011 to 31 December 2016
Number of records:	2,118,143 records [56]
Hereafter referred to as:	Deaths
Population coverage:	The study dataset contains A&E2 data for all HL1 and PREVENT1 households, and for the matched pairs from the most and least deprived cohorts.
Variables
Arrival_Date
Board_of_Treatment ( NHS Health Board)
Age
Gender

Inpatients and Day Cases ( SMR01)

A day case is a patient who has an elective admission to a specialty for clinical care, and sees a professional and requires supervised recovery in the place of treatment. The patient is not expected to, and does not, remain overnight. An inpatient is a patient who occupies, or is expected to occupy, an available staffed bed in a hospital for one or more nights. Inpatients also include all those admitted as an emergency or urgent case. An inpatient or day case admission marks the start of the episode. The patient undergoes the full admission procedure and is accepted by the hospital, the specialty and the consultant for care. For more information, see the General Acute Inpatient and Day Case (SMR01) webpage.

Population Data

Data Controller: NHS National Services Scotland
A record represents: An inpatient or day case episode
Data collection start: 1960 onwards (computerised from 1968)
Frequency of collection: Continuously
Number of annual records: ~1,400,000
Geography coverage: All residents in Scotland that receive care in hospital and general acute specialities.
Population coverage: All Inpatients/ Day cases in Scotland

Study Data

Data period: 1 April 2002 to 31 March 2015
Number of records: 2,266,144 records
Hereafter referred to as: Acute Admissions
Population coverage: The study dataset contains SMR01 data for all HL1 and PREVENT1 households, and for the matched pairs from the most and least deprived cohorts.

Variables:

• HBtreat_current_data ( NHS Health Board)
• Age_in_years
• Sex
• Admission_date (used to determine an overlap with a homelessness episode)
• Discharge_date (used to determine an overlap with a homelessness episode)
• Length_of_stay
• Conditions relating to the admission (1 if related):
  • Alcohol
  • Drugs
  • IorP (Injury and Poisonings)
  • MH (Mental and behavioural Disorders)
  • Resp (Diseases of the Respiratory System)
  • Other. (If an admission has one of the above conditions then this is set to FALSE, otherwise TRUE).

Notes: The study assumes that the health conditions relate directly to the person themselves, even though this may not always be the case. For example, a person hospitalized by a drunk driver would be an alcohol-related admission, but the patient does not necessarily have an alcohol dependence issue themselves. See section 4.3 of the original PBPP application for how the condition flags were assigned using ICD10 codes.

Outpatient data ( SMR00)

The Outpatients ( SMR00) dataset collects episode level data from patients on new and follow up appointments at outpatient clinics in all specialities (except A&E and Genito-Urinary Medicine). For more information, see the Outpatient appointments and Attendances (SMR00) webpage.

Population Data

Data Controller: NHS National Services Scotland
A record represents: An outpatient appointment/attendance
Data collection start: Returns started in 1990s and routinely available from 1997 onwards.
Frequency of collection: Data supplied continually, with a six week submission target.
Number of annual records: ~4,400,000 (1.6 million new outpatients and 2.8 million return outpatients).
Geography coverage: All people offered a new or follow up outpatient appointment at a Scottish NHS hospital.
Population coverage: New outpatient appointments (consultant led clinics). The submission of all return attendances is mandatory regardless of whether or not a procedure is performed.

Study Data

Data period: 1 April 2002 to 31 March 2015
Number of records: 9,014,864 records
Hereafter referred to as: Outpatient attendances
Population coverage: The study dataset contains SMR00 data for all HL1 and PREVENT1 households, and for the matched pairs from the most and least deprived cohorts.

Variables:

• Age_in_years
• Sex
• Clinic_date (used to determine an overlap with a homelessness episode)
• Clinic_attendance (categories ; 1 = 'Patient was seen', 5 = 'Patient attended but was not seen ( CNW: Could Not Wait)', and 8 = 'Patient did not attend and gave no prior warning ( DNA))
• Referral_Source
• GPPRAC_Currentdate ( NHS Board Code)

Prescribing Information System

The Prescribing Information System ( PIS) is the definitive data source for all prescribing relating to all medicines and their costs that are prescribed and dispensed in the community in Scotland. The information is supplied by Practitioner & Counter Fraud Services Division (P& CFS) who is responsible for the processing and pricing of all prescriptions dispensed in Scotland. These data are augmented with information on prescriptions written in Scotland that were dispensed elsewhere in the United Kingdom. Note that prescriptions dispensed within hospitals are not included. For more information, see the Prescribing Information System webpage.

Population Data

Data Controller: NHS National Services Scotland
A record represents: A prescription
Data collection start: From April 1993 onwards
Frequency of collection: Monthly
Number of annual records: Around 100 million data items are loaded per annum
Geography coverage: Scotland
Population coverage: Every prescription dispensed in the community.

Study Data

Data period: 14 January 2009 to 31 March 2015. The study was that CHI linkage may not have been possible with data prior to 2009
Number of records: 9,488,022 records
Hereafter referred to as: Prescriptions
Population coverage: The study dataset contains PIS data for all HL1 and PREVENT1 households, and for the matched pairs from the most and least deprived cohorts.

Variables:

• PrescDate (Prescribed date)
• BNFSubSectionCode (British National Formulary ( BNF) Subsection Code)
• Prescribed drugs flag for:
• Mental health conditions ( BNF Subsections - Hypnotics and Anxiolytics (4.1), Psychoses (4.2), Antidepressants (4.3) and Dementia (4.11))
• Alcohol Dependence ( BNF Subsections: 4.10.1)
• Opioid Dependence ( BNF Subsections: 4.10.3)
• Tuberculosis ( BNF Subsection 5.1.9)

Notes : where comparable data are available, the prescriptions dataset for the 1.3 million people in the cohort for the above conditions typically accounts for fewer than 2% of all prescriptions across Scotland in any given year.

Mental Health Inpatient and Day Case ( SMR04)

The Mental Health Inpatient and Day Case dataset ( SMR04) collects episode level data on patients who are receiving care at psychiatric hospitals at the point of both admission and discharge. The majority of mental health episodes will be dealt with first by GPs and then by Community Health Workers. Only the most acute cases will appear in SMR04. For more information, see the Mental Health Inpatient and Day Case webpage.

Population Data

Data Controller: NHS National Services Scotland
A record represents: A mental health inpatient or day case episode
Data collection start: Collected since the 1960s. Available from 1981 onwards. Routinely available from 1997
Frequency of collection: Continuously
Number of annual records: Approximately 21,000 records per year
Geography coverage: The Mental Health Inpatient and Day Case dataset covers everyone admitted to psychiatric hospitals in Scotland (non-residents of Scotland as well as residents)
Population coverage: All mental health hospital inpatients/day cases in Scotland.

Study Data

Data period: 1 April 2002 to 31 March 2015
Number of records: 100,055 records
Hereafter referred to as: Mental Health Admissions
Population coverage: The study dataset contains SMR04 data for all HL1 and PREVENT1 households, and for the matched pairs from the most and least deprived cohorts.

Variables:

• HBtreat_current_data ( NHS Health Board)
• Age_in_years
• Sex
• Admission_date (used to determine an overlap with a homelessness episode)
• Admission_referral_from (source of referral)
• Discharge_date (used to determine an overlap with a homelessness episode)
• Length_of_stay (days)
• Conditions relating to the admission (1 if related, based on main, second and third admission diagnosis):
  • Alcohol
  • Drugs
  • IorP (Injury and Poisonings)
  • MH (Mental and behavioural Disorders)
  • Resp (Diseases of the Respiratory System)
  • Other. (If an admission has one of the above conditions then this is set to FALSE, otherwise TRUE).

Scottish Drug Misuse Database ( SDMD) ( SMR24 and SMR25a)

The Scottish Drug Misuse Database ( SDMD) offers a profile of the misuse of drugs based on information about drug clients seen at a broad range of services across Scotland. Services contributing to the Database consist mainly of specialist drug services, general practitioners and prison drug services. For more information, see the Scottish Drug Misuse Database webpage.

Population Data

Data Controller: NHS National Services Scotland
A record represents: A drugs misuse assessment
Data collection start: A variety of data has been collected from 1996 onwards. See Technical Guidance for more information. April 2006, SMR25a (covering new assessments) were introduced to reflect the need for more in depth and focussed information from clients.
Frequency of collection: Continuously through web submissions, electronic uploads of batch files of data collected from local systems and formerly on paper forms.
Number of annual records: Approximately 12,000 new assessments and 13,000+ follow-up assessments per year.
Geography coverage: Scotland
Population coverage: Drug users

Study Data

Data period SMR24: 1 April 2002 to 31 March 2006
Data period SMR25a: 1 April 2006 to 31 March 2015
Number of records: 89,281 records
Hereafter referred to as: SDMD
Population coverage: The study dataset contains SDMD data for all HL1 and PREVENT1 households, and for the matched pairs from the most and least deprived cohorts.

Variables:

• Date_assessment
• For PRESCRIPTION DRUGS PROFILE, 6 different drugs are recorded and for each one there is drug code, drug name, daily dosage and measure.
• For ILLICIT DRUGS PROFILE, 8 different drugs are recorded and for each on there is drug code, drug name, main route, other route, how often, daily quantity and measure.

Note: drugs captured in the illicit fields relate to the drugs taken inappropriately that have resulted in the misuse episode. Drugs captured in the prescription fields relate to the treatment of addiction.

National Records of Scotland – Deaths

The NRS produces general publications for Vital Events, specifically, for births, deaths and marriages in Scotland. The dataset covers all death occurring in Scotland and is collected weekly. For more information, see the National Records of Scotland (NRS) – Deaths Data webpage.

Population Data

Data Controller: National Records of Scotland
A record represents: A death event
Data collection start: 1974
Frequency of collection: Weekly
Number of annual records: Approximately 55,000 deaths registered annually
Geography coverage: Scotland
Population coverage: All Deaths occurring in Scotland

Study Data

Data period: 1 April 2002 to 31 March 2015
Number of records: 23,718 records
Hereafter referred to as: Deaths
Population coverage: The study dataset contains all death data for all H2H individuals, and for the matched pairs from the most and least deprived cohorts.

Variables:

• Date_of_death
• Primary_cause_of_death ( ICD10 ^[57] )
• Secondary_cause_of_death (0–9)

Notes: The ICD10 code set allows for more than 14,400 different codes to describe diseases and related health problems. The study mapped the ICD10 code for the primary cause of death to the following groupings:

• alcohol related conditions ^[58] ;
• drug related conditions ^[59] ;
• mental and behavioural disorders ^[60] ;
• diseases of the respiratory system ^[61] ;
• Neoplasms (Cancer) ^[62] ;
• Diseases of the circulatory system (Heart Disease and Strokes) ^[63] ;
• Intentional Self-Harm ^[64] ;
• Assault ^[65] ;
• Diseases of the digestive system ^[66] , and;
• All other deaths.

There were no deaths recorded in the dataset with ICD10 codes beginning with S or T (Injury of Poisonings) ^[67] .