Measures of Disease Occurrence

Kitty J. Jager¹, Carmine Zoccali², Reinhard Kramar³ and Friedo W. Dekker1,4

1 ERA–EDTA Registry, Dept. of Medical Informatics, Academic Medical Center, Amsterdam, The Netherlands

2 CNR–IBIM Clinical Epidemiology and Pathophysiology of Renal Diseases and Hypertension, Renal and Transplantation Unit, Ospedali Riuniti, 89125 Reggio Cal., Italy

3 Austrian Dialysis and Transplant Registry (OEDTR), General Hospital of Wels, Wels, Austria

4 Department of Clinical Epidemiology, Leiden University Medical Centre, Leiden, The Netherlands

Kidney International: Series on epidemiology

Epidemiology

• The study of the occurrence of disease

• Centers of Disease Control and Prevention (CDC):

Epidemiology is the study of the distribution and determinants of health-related states or events in specified populations, and the application of this study to the control of health problems1

• Measures of disease occurrence to be used for analysis depend on the aim of the study

• Prevention, etiology, risk factors new cases Incidence

• Burden of disease, planning of health care existing cases Prevalence

1 http://www.cdc.gov/reproductivehealth/EpiGlossary/glossary.htm

Incidence - Definition of Terms

Concept Definition Formula

Risk Probability of developing disease No of subjects developing disease during a time period

No of subjects followed for the time period

Incidence rate Ratio of the number of cases to the time at risk for disease

No of subjects developing disease

Total time experienced for the subjects followed

Incidence - Risk

Risk (synonyms: cumulative incidence, incidence proportion)

• always necessary to define a time period to which the risk applies

• assumes that subjects are followed for the entire time period

2 Puliyanda DP, Stablein DM, Dharnidharka VR. American Journal of Transplantation 2007; 7: 1–5

Example 1 – Risk of hospitalization for bacterial infection

The paper of Puliyanda et al.2 describes a cohort of 3106 children during the first 2 years post-renal transplantation. One of the purposes of the study was to determine the risk of hospitalization for bacterial infection in the first two years after renal transplantation. 164 children lost their grafts in the first 6 months after transplantation. 687 patients were hospitalized for bacterial infection.

Incidence - Risk

Risk = 687 / 3106 = 22.1%

but …. 164 lost their grafts for other reasons and therefore were not able anymore to develop the event of interest!

In case the follow-up time is long, there are problems with ‘risk’ as a measure of disease occurrence, due to loss to follow-up and competing risk

2 Puliyanda DP, Stablein DM, Dharnidharka VR. American Journal of Transplantation 2007; 7: 1–5

Example 1 – Risk of hospitalization for bacterial infection

The paper of Puliyanda et al.2 describes a cohort of 3106 children during the first 2 years post-renal transplantation. One of the purposes of the study was to determine the risk of hospitalization for bacterial infection in the first two years after renal transplantation. 164 children lost their grafts in the first 6 months after transplantation. 687 patients were hospitalized for bacterial infection.

Incidence – Incidence rate

Incidence rate (synonyms: incidence density, hazard, force of morbidity/

mortality)

• not required for every study subject to complete the entire risk period, as only ‘time at risk’ is taken into account

• this makes the incidence rate very useful in cases where subjects may or may not be at risk for the event of interest for particular periods of time

Example 2 – Incidence rate of peritonitis requiring hospitalization in CAPD patients

Suppose we would have four episodes in 17 patients. We can only calculate the incidence rate if we know the total time at risk (= total time on CAPD).

Incidence – Incidence rate

• In total 144 months at risk

• 4 peritonitis episodes

• Incidence rate =

4/144 = 0.028/patient month

=• 4/12 = 0.33 /patient year

• Incidence rate = reciprocal of the waiting time, i.e. the average time until an event occurs

• Average waiting time for a peritonitis requiring hospitalization to occur in a CAPD patient is 1/ 0.33 = 3 years

died

newly started PD

died

01-01-2004 31-12-2004

Patients

123456789

1011121314151617

transfer to HD

transfer to HD

newly started PD

newly started PD

newly started PD

transfer from HD

Months

Patient months

48

125

12122

12724

126

12101212

1440 1 2 3 4 5 6 7 8 9 10 11 12

died

newly started PD

died

01-01-2004 31-12-2004

Patients

123456789

1011121314151617

transfer to HD

transfer to HD

newly started PD

newly started PD

newly started PD

transfer from HD

Months

Patient months

48

125

12122

12724

126

12101212

1440 1 2 3 4 5 6 7 8 9 10 11 12

died

newly started PD

died

01-01-2004 31-12-2004

Patients

123456789

1011121314151617

transfer to HD

transfer to HD

newly started PD

newly started PD

newly started PD

transfer from HD

Months

Patient months

48

125

12122

12724

126

12101212

1440 1 2 3 4 5 6 7 8 9 10 11 12

newly started PD

died

01-01-2004 31-12-2004

Patients

123456789

1011121314151617

transfer to HD

transfer to HD

newly started PD

newly started PD

newly started PD

transfer from HD

Months

Patient months

48

125

12122

12724

126

12101212

1440 1 2 3 4 5 6 7 8 9 10 11 120 1 2 3 4 5 6 7 8 9 10 11 120 1 2 3 4 5 6 7 8 9 10 11 120 1 2 3 4 5 6 7 8 9 10 11 12

Incidence – Risk vs Incidence rate

• Over short periods of time risk and incidence rate will be similar, as there will be little loss to follow-up and competing risk will only play a minor role

• Therefore, in-hospital mortality for acute renal failure is commonly expressed as a risk, whereas mortality on dialysis for end-stage renal disease is usually expressed as a rate

A comparison of properties of risk and incidence rate (adapted from [3])

Property Risk Incidence Rate

Range 0 – 1 (0 – 100%) 0 - infinity

Units None 1 / time

Interpretation Probability Reciprocal of waiting time

3 Rothman KJ. Epidemiology: an introduction. Oxford University Press 2002

Prevalence - Definition of Terms

• Prevalence is a measure of disease status: it deals with existing cases of disease

Concept Definition Formula

Point prevalence Proportion of people in a population having disease at a particular point in time

No of subjects having disease at a particular point in time

Total no of subjects in the population

Period prevalence Proportion of people in a population having disease over a period of time

No of subjects with disease at the start of the period

+ no of subjects developing disease over the time period

Total no of subjects in the population

Prevalence – Point Prevalence vs Period Prevalence

• Prevalence is a measure of disease status: it deals with existing cases of disease

Example 3 – Prevalence of ESRD due to glomerulonephritis

At the beginning of 2005, 20 patients out of the 80 patients in our dialysis centre had ESRD due to glomerulonephritis (GN). During 2005 there were 2 new patients with the same condition taken into dialysis.

Point prevalence of ESRD due to GN at January 1, 2005: 20 / 80 = 0.25

Period prevalence in 2005 = 22 / 80 = 0.275

Prevalence – Dependence on Incidence rate and Disease Duration

• Prevalence = Incidence rate x Average disease duration

Incidence rate

Survival =

Incidence rate =

Survival

Incidence rate

Survival

Prevalence

Prevalence

Prevalence

Prevalence

• Burden of disease in a population (e.g. costs, life expectancy, morbidity)

• Helps to determine where investment in health should be targeted

Example 4 – Prevalence: economic burden of mineral regulating therapy in dialysis patients

Lorenzo et al. did a study among 1312 haemodialysis patients from 6 centres in Spain4, which was estimated to represent almost 10% of all haemodialysis patients in Spain. They performed a cost analysis to evaluate the economic burden of mineral regulating therapy in this patient group. It turned out that on average the cost of this specific therapy was 1,68 Euro per patient per day.

Costs of mineral regulating therapy: 10 x 1,312 x 365 x 1.68 = 8,045,184 Euro

4 Lorenzo V, Martin-Malo A, Perez-Garcia R, et al. Nephrol Dial Transplant 2006;21:459-465