The study aims of the feasibility is to determine the number of patients with a diagnosis of Pyruvate kinase deficiency (PKD) and their date of diagnosis in primary care data linked with Hospital episode statistics (HES) data. The availability of information on healthcare resources involved with PKD such as hospital visits and treatment received will be accessed. Clinical characteristics will also be accessed such as age, sex and other factors that are associated with PKD such as history of splenectomy and blood transfusions.
All patients with a diagnosis of PKD in their case records will be included. The feasibility will be used to determine the level of recording of the factors such as those mentioned above in the data source to help better define the point a patient is diagnosed with PKD
PKD is a rare disease and its true prevalence in the UK is unknown, likely due to misdiagnosis of mildly affected patients. By describing Healthcare resource utilisation (HCRU) and patient characteristics of these patients this study may help to better define patients and to understand HCRU associated with PKD to better inform treatment areas.
Results from this feasibility will be used to carry out the main study objective of describe the burden of illness in PKD patients compared to non PKD patients and to describe characteristics associated with PKD.
Objectives
1. To determine number of patients with a diagnosis of PKD in CPRD Gold/Aurum linked with HES
2. To validate the accuracy of the assumption of a diagnosis of PKD in CPRD Gold and Aurum for patients with a READ code of “Pyruvate kinase” or “Pyruvate kinase screening test” through data enhancement
3. To collect more detailed clinical information on patients in CPRD Gold and Aurum with a READ code of “Haemolytic anaemia due to pyruvate kinase deficiency”, “Pyruvate kinase” or “Pyruvate kinase screening test” through data enhancement
4. To describe the completeness and availability of study variables required for the full study
5. To describe the demographic and high-level clinical characteristics (excluding detailed description of co-morbidities) as well as follow-up (pre- and post-index) of the patients with PKD and validated in Objectives 1 and 2
The PKD cases based on READ codes will be described by number of patients and date. For patients with a READ code for “Pyruvate kinase” or “Pyruvate kinase screening test” only, the test events included under “Pyruvate kinase screening test” will be explored. Where these include a negative test result, these patients will not be further considered. Patients with a positive test result or without a test result will be included in the data enhancement process.
For the cases described above, high-level patient characteristics as well as pre- and post-index follow-up time will be described.
The study variables for the full study (see tables 2, 3 [with the exception of comorbidity variables], 4, 5 and 6) will be described by number of patients and period of measurement. Based on these results, the total of patients available in different potential HCRU observation periods will be described to inform decisions on the HCRU observation period for the full study.
C. Health Outcomes to be Measured
The following variables will be assessed for completeness and availability to further define the study population and variables such as index date and first entry in to the database to better determine the observation window for HCRU analysis in the full study. Results from the feasibility will be used to help carry out the main study of measuring and describing healthcare resource utilisation of PKD patients compared to matched controls and to describe detailed clinical and disease characteristics of PKD patients.
For all variables listed below CPRD small number rules will be applied to all study outputs to avoid identification of individual patients, this means any number <7 will be masked, along with the next lowest number in the column. To further avoid identification numbers will also be rounded to the nearest 5, note row/column totals will be calculated and then rounded so rounded values may differ from rounded totals.
Limited stratification will be used in the feasibility to help prevent low numbers.
The following read codes will be used to capture Pyruvate kinase disease patients:
Read Code Description
42G5.00 Pyruvate Kinase
D103100 Haemolytic anaemia due to pyruvate kinase deficiently
42G7.00 Pyruvate Kinase Screening test
42G5.11 Red cell pyruvate Kinase test
Table 1: PKD case definition
Certainty of diagnosis Description Operational definition in CPRD Gold and Aurum Timing
Definite PKD diagnosis based on READ code READ code of “Hemolytic anemia due to pyruvate kinase deficiency”:
• Yes or no
• Date of first READ code Structured data (READ code/SNOMED CT/EMIS code) available in data source Within study observation period
Possible PKD diagnosis based on READ code READ code of “Pyruvate kinase”:
• Yes or no
• Date of first READ code
READ code of “Pyruvate kinase screening test”:
• Yes or no
• Positive or Negative test result
• Date of first positive test result Structured data (READ code/SNOMED CT/EMIS code) available in data source
To be explored and confirmed if READ code/SNOMED CT/EMIS code is associated with test results (positive/negative) Within study observation period
Definite PKD diagnosis based on lab test (confirmed by GP) Lab test confirmed PKD diagnosis:
• Yes or no
• Date of first lab test confirmation Data enhancement needed (using GP questionnaire, see Annex 1. Data enhancement questionnaire)
Any time
Table 2: HCRU variables
Variable Description Operational definition in CPRD Gold and Aurum Timing
Primary care attendances Face-to-face visits, home visits, telephone calls:
• Number
• Date
Administrative visits will be excluded (as these are expected to represent administrative updates to the patient file only) As recorded in data source From index date to end of follow-up
Specialist clinic referrals Any specialist clinic referral, outpatient visits:
• Number
• Type of specialist clinic (e.g., haematology)
• Date As recorded in data source, Data enhancement From index date to end of follow-up
Hospitalisations Any inpatient stay, day case visit, A&E visit:
• Number
• Reason for hospitalisation
• Length of stay (dates) As recorded in data source, HES linkage (reason for hospitalisation: primary ICD-10 code), Data enhancement
Reason for hospitalisation is not available in CPRD or HES. In HES: based on the primary ICD-10 code during hospitalization. From index date to end of follow-up
Routine laboratory tests Full blood count (e.g., haematocrit, haemoglobin, platelet count, red blood cell count):
• Number
• Date of test Structured data (READ code/SNOMED CT/EMIS code) available in data source, HES linkage (ICD-10 code) From index date to end of follow-up
Splenectomy Splenectomy:
• Yes or no
• Date of procedure
Code list will be included in the SAP. Structured data (READ code/SNOMED CT/EMIS code) available in data source, HES linkage (ICD-10 code) From index date to end of follow-up
Cholecystectomy Cholecystectomy
• Yes or No
• Date of procedure
Code list will be included in the SAP. Structured data (READ code/SNOMED CT/EMIS code) available in data source, HES linkage (ICD-10 code) From index date to end of follow-up
Blood transfusion Blood transfusion:
• Number
• Dates of blood transfusion Structured data (READ code/SNOMED CT/EMIS code) available in data source, HES linkage (ICD-10 code), Data enhancement From index date to end of follow-up
Venesection Venesection:
• Number
• Dates of venesection procedures Structured data (READ code/SNOMED CT/EMIS code) available in data source, HES linkage (ICD-10 code), Data enhancement From index date to end of follow-up
Primary care prescriptions Primary care prescriptions that will be considered are those for antibiotics, anticoagulants, antidepressants and pneumococcal vaccination:
• Active substance
• Dose
• Formulation
• Dates of prescription
Code list will be included in the SAP. Structured data (prescribing codes) available in data source From index date to end of follow-up
Iron chelation therapy Iron chelation therapy (see Table 1 for agents to be considered):
• Active substance
• Dose
• Formulation
• Dates of prescription Structured data (prescribing codes) available in data source From index date to end of follow-up
Stem Cell Transplant Stem cell Transplant:
• Yes or No
• Date Structured data (READ code/SNOMED CT/EMIS code) available in data source, HES linkage (ICD-10 code), Data enhancement From index date to end of follow-up
Table 3: Patient clinical characteristics variables
Variable Description Operational definition in CPRD Gold and Aurum Timing
Age Age in years (continuous variable) As recorded in data source At index date or start of the HCRU observation period (to be defined once the HCRU observation period has been confirmed)
Sex Sex:
• Female or male As recorded in data source Any time
Body mass index (BMI) BMI in kg/m2.
If BMI is not directly available, it will be calculated from height and weight:
BMI = weight (in kg) / [height (in m) x height (in m)] As recorded in data source Closest recording to index date or to start of the HCRU observation period (to be defined once the HCRU observation period has been confirmed)
Socioeconomic status Index of Multiple Deprivation (IMD) score based on the patient’s postcode:
• Categories will be defined in the SAP As recorded in data source At index date or start of the HCRU observation period (to be defined once the HCRU observation period has been confirmed)
Region Region in the UK based on the patient’s/practice’s postcode:
• Categories will be defined in the SAP As recorded in data source At index date or start of the HCRU observation period (to be defined once the HCRU observation period has been confirmed)
Pre- and post-index follow-up time Pre-index follow-up time will be calculated as time from date of enrolment in CPRD to index
Post-index follow-up time will be calculated as time from index date to end of follow-up As recorded in data source Within the study observation period
Iron overload syndrome Iron overload syndrome:
• Yes or no
Code list will be included in the SAP. Structured data (READ code/SNOMED CT/EMIS code) available in data source Within the study observation period
Osteopenia Osteopenia:
• Yes or no
Code list will be included in the SAP. Structured data (READ code/SNOMED CT/EMIS code) available in data source Within the study observation period
Fragility fracture Fragility fracture:
• Yes or no
Code list will be included in the SAP. Structured data (READ code/SNOMED CT/EMIS code) available in data source Within the study observation period
Depression Depression:
• Yes or no
Code list will be included in the SAP. Structured data (READ code/SNOMED CT/EMIS code) available in data source Within the study observation period
Skin ulceration Skin ulceration:
• Yes or no
Code list will be included in the SAP. Structured data (READ code/SNOMED CT/EMIS code) available in data source Within the study observation period
Cirrhosis Cirrhosis:
• Yes or no
Code list will be included in the SAP. Structured data (READ code/SNOMED CT/EMIS code) available in data source Within the study observation period
Pulmonary hypertension Pulmonary hypertension:
• Yes or no
Code list will be included in the SAP. Structured data (READ code/SNOMED CT/EMIS code) available in data source Within the study observation period
Gallstones History of Gallstones
• Yes or No
• Number of episodes
Date of episodes Structured data (READ code/SNOMED CT/EMIS code) available in data source Within the study observation period
Venous thrombosis Venous thrombosis:
• Yes or no
Code list will be included in the SAP. Structured data (READ code/SNOMED CT/EMIS code) available in data source Within the study observation period
Splenectomy Splenectomy:
• Yes or no
• Date of procedure
Code list will be included in the SAP. Structured data (READ code/SNOMED CT/EMIS code) available in data source, HES linkage (ICD-10 code) Within the study observation period
Prophylactic antibiotics prophylactic antibiotic use:
• Yes or No
Code list will be included in the SAP.
All prescription for antibiotics will be captured
Structured data (READ code/SNOMED CT/EMIS code) available in data source Within the study observation period
Death Death:
• Yes or no
• Date of death
• Cause of death As recorded in data source Within the study observation period
Table 4: PKD characteristics variables
Variable Description Operational definition in CPRD Gold and Aurum Timing
Blood transfusion Blood transfusion:
• Number
• Dates of blood transfusion HES linkage, Data enhancement Within the study observation period
Iron overload READ code for iron overload.
Code list will be included in the SAP.
If a READ code for iron overload is not available, it can be inferred by prescription for iron chelation therapy. The case definition of iron overload will be defined in light of the findings from the feasibility study. Structured data (READ code/SNOMED CT/EMIS code, prescribing codes) available in data source Within the study observation period
Iron chelation therapy Iron chelation therapy (see Table 1 for agents to be considered):
• Active substance
• Dose
• Formulation
• Dates of prescription Structured data (prescribing codes) available in data source From index date to end of follow-up
Time from PKD diagnosis to first initiation of iron chelation therapy Time from PKD diagnosis to first initiation of iron chelation therapy (in patients receiving iron chelation therapy) Structured data (prescribing codes) available in data source From index date to end of follow-up
Table 6: Bloodborne virus infection variables
Variable Description Operational definition in CPRD Gold and Aurum Timing
Bloodborne virus infection (Hepatitis B, Hepatitis C) Bloodborne virus infection:
• Yes or No
• Date of virus infection
Code lists will be included in the SAP. Structured data (READ code/SNOMED CT/EMIS code) available in data source From index date to end of follow-up
Table 6: Subgroup variables
Variable Description Operational definition in CPRD Gold and Aurum Timing
Iron chelation Iron chelation:
• Yes or No Structured data (prescribing codes) available in data source Within study observation period
Age Age categories:
• <18 years or ≥18 years As recorded in data source At index date or start of the HCRU observation period (to be defined once the HCRU observation period has been confirmed)
Joseph Kim - Chief Investigator - IQVIA Ltd ( UK )
Peter McMahon - Corresponding Applicant - IQVIA Ltd ( UK )
Ben Bray - Collaborator - IQVIA Ltd ( UK )
Jessica Lundbom - Collaborator - IQVIA Ltd ( UK )
Joshua Warden - Collaborator - IQVIA World Publications Ltd.
Pesheya Doubleday - Collaborator - IQVIA World Publications Ltd.
Rachel Tham - Collaborator - IQVIA Ltd ( UK )
HES Accident and Emergency;HES Admitted Patient Care;ONS Death Registration Data