Ngā Tapuae me ngā Raraunga: Methods and data sources

ICD-10 codes

Table A1.1: Amenable mortality codes
Condition ICD-10
Pulmonary tuberculosis A15
Meningococcal disease A39
Pneumococcal disease J13, A40.3, G00.1
HIV/AIDS B20–B24
Cancer
Stomach C16
Rectum C19–C21
Melanoma C43
Female breast C50
Cervix C53
Testis C62
Prostate C61
Thyroid C73
Bone and cartilage C40–C41
Hodgkins C81
Acute lymphocytic leukaemia C91.0
Complications of pregnancy O01–O99
Complications of the perinatal period P02–P94
Congenital heart disease (subset) Q21
Diabetes E10–E14
Valvular heart disease I01, I05–I09, I33–I37
Hypertensive diseases I10–I15
Coronary disease I20–I25
Heart failure I50
Cerebrovascular diseases I60–I69
Renal failure N17–N18
Pulmonary embolism I26
COPD J42
Asthma J45–J46
Peptic ulcer disease K25–K26
Cholelithiasis K80
Suicide X60–X84
Road traffic accidents V01–V79, V87, V89, V99
Falls (#NOF) S72
Burns T20–T31
Adverse health care events (subset) T80–T88

# fractured neck of femur

Table A1.2: Ambulatory-sensitive hospitalisation codes
Condition ICD-10
Angina and chest pain I20, R07.2–R07.4
Asthma J45–J46
Bronchiectasis J47
Cellulitis H00.0, H01.0, J34.0, L01–L04, L08, L98.0
Cervical cancer C53
Congestive heart failure I50, J81
Constipation K59.0
Dental conditions K02, K04, K05
Dermatitis & eczema L20–L30
Diabetes E10–E14, E162
Epilepsy G40–G41, O15, R56.0, R56.8
Gastroenteritis/dehydration A02–A09, R11
GORD (Gastro-oesphageal reflux disease) K21
Hypertensive disease I10–I15, I67.4
Kidney/urinary infection N10, N12, N13.6, N30.9, N39.0
Myocardial infarction I21–I23, I24.1
Nutrition deficiency and anaemia D50–D53, E40–E46, E50–E64, M83.3*
Other ischaemic heart disease I240, I24.8,I24.9, I25
Peptic ulcer K25–K28
Respiratory infections – Pneumonia J13–J16, J18
Rheumatic fever/heart disease I00–I02, I05–I09
Sexually transmitted infections A50–A59,A60, A63, A64, I980, M02.3, M03.1, M73.0, M73.1, N29.0, N34.1
Stroke I61, I63–I66
Upper respiratory tract and ENT infections J00–J04, J06, H65–H67
Vaccine-preventable disease – Meningitis, Whooping cough, Hepatitis B, Pneumococcal disease, Other A33–A37, A40.3, A80, B16, B18
Vaccine-preventable disease – MMR B05, B06,B26, M01.4, P35.0**

*Adult only (15+ years)
**All ages

Table A1.3: ICD-10 codes used in this report
Condition ICD-10
Total cardiovascular disease I00–I99
Ischaemic heart disease I20–I25
Other forms of heart disease I30–I52
Heart failure I50
Total stroke I60–I69
Rheumatic heart disease I05–I09
Pneumonia J12­–J18
Chronic obstructive pulmonary disease (COPD) J40–J44
All revascularisation (CABG and angioplasty) heart disease procedures 3530400, 3850500, 9022100, 3530500, 3531000, 3531002, 3849700, 3849701, 3849702, 3849703, 3850000, 3850300, 3849704
Diabetes E10–E14
Diabetes complications with renal failure E102, E112, E122, E132, E142
Lower limb amputation with concurrent diabetes E10–E14 together with
4433800, 4435800, 9055700, 4436100, 4436400, 4436401, 4436101, 4437000, 5023600, 4437300, 5023300, 4436700, 5023602, 4436701, 4436702
All cancers C00–C97
Stomach cancer C16
Colorectal cancer C18–C21
Liver cancer C22
Lung cancer C33–C34
Prostate cancer C61
Breast cancer (female only) C50
Uterine cancer C54–C55
Cervical cancer C53
All injuries V01–Y98
Unintentional injuries (Accidents) V01–X59
Motor vehicle traffic V20–V59
All other transport V60–V99
Falls W00–W19
Machinery W28–W31
Firearms W32–W34
Drownings and submersions W65–W74
Suffocation W75–W84
Fires/hot objects or substances X00–X19
Poisonings X40–X49
Suicide and self-harm X60–X84
Assault and homicide X85–Y09

2001 Census total Māori population

Table A2.1: 2001 Census Māori population
Age group (Years) Number Weighting
0–4 67,404 12.81
5–9 66,186 12.58
10–14 62,838 11.94
15–19 49,587 9.42
20–24 42,153 8.01
25–29 40,218 7.64
30–34 39,231 7.46
35–39 38,412 7.30
40–44 32,832 6.24
45–49 25,101 4.77
50–54 19,335 3.67
55–59 13,740 2.61
60–64 11,424 2.17
65–69 8043 1.53
70–74 5046 0.96
75–79 2736 0.52
80–84 1251 0.24
85+ 699 0.13

Ethnicity: Adjusters for the analysis of hospitalisation data

Background

This appendix describes the method used to create the adjusters used in the analysis of hospital discharge data*. These ethnicity adjusters were created and used to calculate hospitalisation rates in Tatau Kahukura: Māori Health Chart Book 2010, 2nd edition (Ministry of Health 2010d).

High-quality ethnicity data are essential for monitoring health trends by ethnicity. Such data are also needed to provide Māori with high-quality information about Māori health and disparities for planning, for the development and evaluation of policies, and for interventions (Cormack and Harris 2009). However, official health data sets have still been shown to undercount Māori in cancer registrations and hospital admissions, and there is a need to improve ethnicity data in health information systems. The Ministry of Health has ethnicity data protocols for the health and disability sector that outline the procedures that are to be used for the standardised collection, recording and output of ethnicity data for the sector (see Ministry of Health 2004).

According to previous research findings from the New Zealand Census − Mortality Study (NZCMS), the ethnicity records in the death registrations for the years 2001–2004 showed no net undercount of Māori deaths (Fawcett 2008). However, cancer registration data sets in the years 1981–2004 have been shown to undercount Māori cancer registrations (Harris et al 2007; Shaw et al 2009).

In 2009 the methodology used to assign ethnicity to cancer registrations changed. Ethnicity is assigned to cancer registrations by looking at the ethnicity recorded on each of the corresponding death registrations, hospitalisation records and National Health Indexes (NHIs). A cancer registration is automatically assigned the ethnicity(s) on death registrations and NHIs (unless the ethnicity is not stated, or is identified as ‘Other’). In addition, if a particular ethnicity is recorded on at least 20 percent of hospitalisation records, the ethnicity is assigned to the cancer registration.

This means that when there are different ethnic groups on the different source data sets, multiple ethnicities are recorded on the cancer register. This chart book does not adjust for an undercount, so cancer registration rates for Māori could still be underestimated. Further information about the current methodology used to assign ethnicity to cancer registrations can be obtained from the Ministry of Health by emailing data-enquiries@moh.govt.nz.

The ‘ever Māori’ method of ethnicity classification was used in the previous edition of Tatau Kahukura to adjust for the undercount in death records, cancer registration and hospitalisation data**. However, concerns with potential over-counting using this method for more recent time periods has prompted the recommendation that new ethnicity adjusters be developed to address the continued undercount of Māori in hospital discharge data (Harris et al 2007).

Method

Death registration ethnicity was assumed to be a reliable count of Māori ethnicity data. Using encrypted NHIs, public hospital event records were linked to death registrations among those people who had both been admitted to hospital and died in the period 2003–2006. Death records were only available up to 2006, whereas hospitalisation data were available up to 2008. The time period 2003–2006 was chosen because it was the closest period to the period of interest for hospitalisations (2006–2008), and it was wide enough to provide enough data to calculate reliable adjusters. The number of Māori hospitalisations using hospital event ethnicity was compared to the number of Māori hospitalisations using death registration ethnicity. Anyone recorded as Māori (either alone or in combination with another ethnic group or groups) was classified as Māori. Everyone else was classified as non-Māori.

Ratios of Māori hospitalisations (death ethnicity/hospital event ethnicity) by age are presented in Table A3.1 below. A ratio greater than 1 indicates more Māori hospitalisations using death ethnicity for that age group compared with Māori hospitalisations using hospital event ethnicity and therefore suggests an undercount of Māori in the hospitalisation data.

Undercounting of Māori tends to be higher in younger and older age groups. However, the data in younger age groups may be less reliable due to the smaller numbers of deaths, and therefore fewer linkages.

Age-specific smoothed hospital adjusters were created using local regression with the LOESS procedure in SAS (version 9.1, SAS Institute Inc, Cary NC). Smoothing the ratios accounts for the effect of low numbers in younger age groups and the potential unreliability. The smoothed ratios (adjusters) are all above 1 and increase with age.

Table A3.1: Ratios of Māori hospitalisations (death ethnicity/hospital event ethnicity) by age
Age group (Years) Māori recorded at 2003–2006 death registration Māori recorded at 2003–2006 public filtered hospital admission Ratio (death/hospital) Smoothed ratio^ Linked hospital and mortality data 2000–2004 (from Hauora IV)^^
0–4 955 864 1.105 1.027 1.144
5–9 95 128 0.742 1.032 1.084
10–14 272 230 1.183 1.037 1.309
15–19 428 429 0.998 1.041 1.192
20–24 423 376 1.125 1.045 1.132
25–29 350 308 1.136 1.049 1.167
30–34 649 581 1.117 1.053 1.059
35–39 935 919 1.017 1.058 0.999
40–44 1505 1499 1.004 1.064 1.009
45–49 2119 2009 1.055 1.069 1.084
50–54 2769 2605 1.063 1.073 1.068
55–59 3104 2951 1.052 1.078 1.048
60–64 4266 3992 1.069 1.086 1.046
65–69 4121 3939 1.046 1.094 1.040
70–74 3725 3498 1.065 1.102 1.125
75–79 2864 2552 1.122 1.110 1.137
80–84 1760 1578 1.115 1.120 1.153
85+ 1161 951 1.221 1.129 1.161

^The ratios were smoothed using local regression with LOESS procedure in SAS.
^^Robson and Harris 2007

Table A3.1 shows the public hospital adjusters developed for Hauora IV for comparison. The pattern and magnitude of the ratios for this edition of Tatau Kahukura are generally similar to those found in Hauora IV.

The standard error for the smoothed adjusters was also calculated. This standard error was incorporated into the 95 percent CIs for the hospitalisation rates and ratios.

Summary

For the purposes of this chart book, these hospital adjusters are likely to improve the counts for Māori hospitalisations, assuming that death registration data records ethnicity data accurately for Māori.

*This linkage method was developed in Hauora: Māori Standards of Health IV (Robson and Harris 2007), and we would like to acknowledge the authors’ contribution to this report.
**For information on the ‘ever Māori’ method, please see Appendix 3 in the first edition of Tatau Kahukura: Māori Health Chart Book (Ministry of Health 2006).

Back to top