Background
Multimorbidity (MM) is the presence of 2 or more long-term health conditions in a single
individual. It impacts an individual’s quality of life, mental health and wellbeing, daily
function, and often results in greater healthcare utilisation the more co-existing
conditions they have (1-4). MM is a big challenge facing the NHS, especially given
England’s ageing population, with an estimated two-thirds of individuals aged 65 and
over having 2 or more long-term conditions (5-6). Yet, little is known about the resource
use of these patients despite being the group with the largest impact on the NHS and
with the worst health outcomes (7). Existing evidence focuses on specific health
conditions and their interactions with other conditions using different methodologies,
making comparisons across different conditions difficult.
This work has empirically assessed the impact of multi-morbidity on NHS and social
care costs. With the aim of answering the question: is the impact of developing a
condition on health and social care costs greater for someone with no prior conditions,
or for someone with an existing condition. If patients have multiple conditions, there
may be some economies of scale involved with treatment, for example they may be
able to discuss multiple queries during a single GP appointment, or in some cases the
treatment provided will address multiple conditions. However, treating patients with
multi-morbidities could theoretically also be more expensive than treating 2 conditions
separately, as patients may be more likely to experience complications.
Methodology
This work considered the individual cost of 11 health conditions with high prevalence in
the English population and their most common interactions. These were: chronic
obstructive pulmonary disease (COPD), diabetes (types 1 and 2), lung cancer, breast
cancer, coronary heart disease (CHD), stroke, hypertension, dementia, liver disease,
depression and colorectal cancer.
This project had 2 components: a literature review and an empirical estimation of the
costs associated with MM. The literature review was used to inform and establish the
methodology used in the empirical estimation.
The empirical estimation used data on primary healthcare, secondary healthcare, and
prescriptions usage from 2015 to estimate annual aggregated healthcare costs per
patient. We assessed the cost impact of MM in a systematic way by applying advanced econometric methods to account for the specificities of the data distribution. Our
methodology allowed us to attribute healthcare costs to specific conditions.
For social care costs, we calculated the estimated costs using 2 different
methodologies.
For the first (preferred) methodology, we used Somerset Symphony data to calculate
the 2014/15 social care costs of patients in South Somerset. This is a dataset that
combines primary healthcare, secondary healthcare, and social care data. We thus
applied the same methodology that was used to calculate primary and secondary
healthcare costs.
For the second methodology, we used the estimated health-related quality of life for
patients with different conditions and combinations of conditions. We then used a
regression (‘line of best fit’) to estimate their probability of requiring social care. Finally,
we used unit cost estimates to arrive at estimated values for the costs of social care for
individuals with different diseases.
What this publication adds
Average ‘cost per case’ estimates for individuals with single conditions or multimorbidities, each calculated based on the average age of patients with the condition or
multi-morbidity of interest.
These average ‘costs per case’ figures are always higher for individuals with multimorbidities than individuals with a single condition, as individuals with multi-morbidities
tend to be older and additional conditions incur additional costs.
We found that the cost of treating an individual with a multimorbidity is not statistically
different than the additive cost of treating 2 individuals, each with one of the conditions,
controlling for age and costs unrelated to the condition. As an illustrative example, if it
costs £200 to treat a patient with depression and £200 to treat a patient with CHD, we
did not find any evidence that it would cost more than £400 to treat a single patient with
both depression and CHD (controlling for age and unrelated disease costs).
In numerous cases, when considering healthcare costs, we have found that multimorbidity is associated with a reduction of the total individual cost compared to the sum
of individual costs of patients. For example, a male patient with diabetes and CHD will
cost between 77% and 78% (depending on the definition of sample prevalence) of the
cost of treating 2 patients, one with diabetes and one with CHD, controlling for age and
unrelated costs. Applying the same methodology for social care costs as for healthcare costs, we did
not find any evidence that multi-morbidity is associated with either an increase or a
reduction in total individual cost compared to the sum of individual costs of patient, for
social care costs. This may be due to the relatively small sample size of the South
Somerset data we used to estimate social care costs.
Applying the alternative methodology for social care costs, which estimated social care
need based on age and quality of life, we estimated higher social care costs than we
found by analysing the South Somerset data. This implies that social care need may be
greater than local authority social costs in South Somerset. This may be due to the
relative affluence of South Somerset, which would limit the proportion of patients
eligible for local authority-funded social care.