Epidemiology: Measuring & Describing Disease 3

Question 1

What is e-Referral Service (e-RS)

Answer 1

E-referral service is a computer system that takes GP referrals and puts them into hospitals.

Question 2

What are the 2 types of e-RS referrals

Answer 2

-Routine referral -2 week wait referral (typically for expecting cancer)

Question 3

count vs rate

Answer 3

Rate is per unit-time and divided by number of people. (accounts for number of people in each population)

Question 4

Why may there be more cancer referrals to outer London hospitals than in inner London

Answer 4

• more elderly people - affluent populations - referred more as have faster and higher access to healthcare

Question 5

Describe this data

• Describe the trend of 2WW respiratory referrals and explain what might be going on.
• Why is the event rate higher in the outer boroughs? You might want to develop three testable explanations.
• Over the period which boroughs in NWL are showing the highest and lowest incidence?

How could standardisation and other epidemiological approaches be employed to better understand the inequalities in referral and access?

Answer 5

respiratory referral count decreased in March 2020 (COVID19) but as lockdown eased, it increased again but still not to original

• elderly population
• affluent areas

Highest: Hillingdon and Lowest: Westminster

allow us to standardise age or compare to national statisitcs

Question 6

What does the Standardised incidence ratio allow us to see when comparing referrals in populations - describe the inequality

Answer 6

the SIR shows higher burden of disease in the inner city boroughs: Hillingdon and Harrow have low SIR

lung cancer incidence, when you adjust for that age distribution difference is higher in London signalling the deprivation and the inequalities

more likely to end up with lung cancer, but they’re less likely to reach out and get treatment for it - the mortality is much higher because people from poorer backgrounds present later, they will often end up with worse treatment and they are more likely to die = inequality

Question 7

How can count, rate and SIR be used

Answer 7

help us see the data in a bigger picture and allow us to make more inferences from the data

reading between the count, rate and SIR can enable us to infer a range of system issues

• Count – that is the number of referrals
• Rate – the number of referrals divided by head of population
• Standardised incidence ratio – the difference between observed and expected values

Question 8

Why may COVID cases be lower in coastal areas than in rural towns

Answer 8

• tourisms
• mask wearing
• compliance of social distancing
• testing ease
• capacity of hospitals
• reporting of cases
• less pollution/air quality
• less crowded (population density)
• underlying health conditions/comorbidities

Question 9

Key points for lecture

Answer 9

Key points

a. Direct and indirect standardisation enable interpretation and comparison of descriptive epidemiology. Crude measures of frequency or risk can be misleading.
b. Demand and need are drivers for service distribution. However, the inverse care law

demonstrates an ongoing mismatch; health services tend to be poorer and more stretched in more deprived areas.

c. Chance, artefact and signal should be considered as explanations for epidemiological findings.

Question 10

Question: Are incidence and incidence rate different measures or are they referring to the same thing? In the tutorial it states they are comparable “Incidence rate (aka. Incidence)”

Answer 10

Answer

Good question - and the real answer is that epidemiologists are often less than clear what they really mean. Theoretically: incidence is the number of new cases. For the purposes of this example, let’s say the incidence is five. Incidence rate meanwhile means that the numerator (five) is divided or averaged across a unit of time: for example, if these five cases took place over two years, then the incidence rate would be 2.5 cases per year.

The reality is that the term ‘incidence’ is often used to mean ‘incidence rate’. And in fact, incidence rate is normally the statistic we’re interested in. You are absolutely right to pick up on this nuance, and it is the type of issues that is commonly highlighted at peer review!

Question 11

Question: I have a question about the epidemiology work that I thought of after the session and was wondering if you could answer. When looking at incidence data, how can you tell if the data is the cumulative incidence or the incidence rate? I was unsure which the NHS referral data we were looking at today - is it safe to just assume any incidence data will always be the incidence rate?

Answer 11

Answer

Remember incidence rate is a rate, and cumulative incidence is a proportion.

The incidence rate should be divided by a named time period. In the case of COVID-19 incidence rate has been typically used as either:

Incidence rate (cumulative) - the time period being “since the start of the pandemic” - note that this is not the same as cumulative incidence (which is a proportion).

Incidence rate (7, 14 or 28 days) - the time period picking out a chunk of time

In terms of NHS referral data, we normally look at incidence with the time period being per week or per month. Weeks are more often used as they are a fixed seven day unit. Months are sometimes more problematic because of the number of weekends and February being shorter than the others. For example we often see lower presentation numbers to emergency departments in February because there are fewer days.