The chronic disease explosion: artificial bang or empirical whimper?BMJ 2016; 352 doi: https://doi.org/10.1136/bmj.i1312 (Published 15 March 2016) Cite this as: BMJ 2016;352:i1312
- Kimberlyn McGrail, associate professor1,
- Ruth Lavergne, postdoctoral fellow2,
- Steven Lewis, president3 4
- 1School of Population and Public Health, University of British Columbia, Vancouver, BC, Canada V6T 1Z3
- 2McGill University Montreal
- 3Access Consulting Saskatoon
- 4Simon Fraser University Burnaby
- Correspondence to: Kimberlyn McGrail
- Accepted 26 February 2016
The World Health Organization defines health as the complete state of physical, mental, and social wellbeing. A recent paper from the Global Burden of Disease study suggests only 4% of the world’s population is now free of disease.1 If true, health is not the norm but an anomalous state, and only a tiny fraction of the world’s population could meet the WHO’s exacting standard. The main culprit, chronic disease, is now so common that healthcare systems are increasingly focused on people with more than one condition. Multimorbidity, recently described as the “most common chronic condition,”2 affects at least half of the population over age 65.3
The effect of chronic conditions, and particularly multimorbidity, on healthcare expenditures is striking. Patients with chronic conditions now account for most consultations in primary care4 and an estimated 84% of total spending in the US.5 Not surprisingly, costs are related to the number of chronic conditions. Compared with people with no chronic conditions, costs are nearly three times greater for people with one condition, five times greater for those with two, and eight times greater for those with three.5
Moreover, recent reports suggest that the number of people with chronic disease is rising faster than previously projected.5 The combination of increasing prevalence and the cost of treatment has led to widespread consensus that dealing with and managing chronic conditions is the single largest challenge facing healthcare systems in the developed world.4
This is a gloomy scenario, but only if reported increases reflect real health problems that can benefit from enhanced diagnosis and treatment. There has been little or no critical assessment of the pace of epidemiological change, the plausibility of proposed explanations for the sharply upward turn in prevalence, or implications for the population’s health.
Figure 1⇓ shows the prevalence of 11 chronic diseases over time derived from data on use of healthcare services in British Columbia, a Canadian province with universal publicly financed hospital and physician services and a population of about 4.5 million.6 7 8 These 11 conditions have been targeted for policy and practice change because of their prevalence or effect on healthcare expenditure.9 As is usual in the chronic disease literature, a condition is counted only when there are at least two diagnoses from physician encounters or one hospital based diagnosis.10 We measured prevalence in four year windows.
That prevalence increases with age is to be expected; what is striking is the rate of increase in age specific rates over time. For example, in 10 years the chronic disease burden in people aged ≥80 rose from an average of 1.4 conditions to 2 conditions, and in the 70-79 age group it rose from 1.2 to 1.5. As fig 2⇓ shows, the increase among the ≥80 age group comes from both a smaller proportion of people having no diagnosed condition (only a fifth in 2012-13) and a much higher proportion having three or more conditions (almost two fifths in 2012-13).
There are three possible explanations for the rapid increases in rates of chronic disease (table 1⇓): a true pandemic of chronic disease, especially pronounced in elderly people; earlier or more vigilant recording of existing and known diagnoses; or “diagnostic creep”—people are now diagnosed with a chronic disease who would not have been so a decade or two ago, creating a larger patient population with ongoing care needs. We consider each of these potential explanations in turn.
Is health declining?
At an individual level we know that health generally declines as we get older. This is true at a population level as well, though the pace of change in populations is much slower. However, fig 1⇑ shows an age specific increase over time.
A larger proportion of each birth cohort is living to older ages, and in some cases this will be because of healthcare interventions. A heart attack that used to be lethal might now be treated successfully, leaving an individual alive but with a new chronic condition. This would mean that at least some proportion of the population is, indeed, older and sicker. These effects, however, would have to be very large to be responsible for the trends shown in fig 1⇑; it is implausible that they explain the entire increase over such a short period.
Population risk factors are also changing over time. Some risk factors, like obesity, are more prevalent while others, like smoking, are less prevalent. More people are living with diabetes but fewer with heart disease. These opposing trends are a challenge to intuitive assessments about the influence of risk factors on chronic disease. For example, microsimulation modelling predicts heart disease will continue to decline because the increasing trends in obesity are more than offset by decreases in smoking.11
Declining health is also counter to population reports on self rated health status and functional ability. Fig 3⇓ shows that self reported health status is improving despite increasing chronic disease. One possible explanation is that changes in diagnosis and treatment lessen the impact of chronic disease on self reported health. Another is that people who are diagnosed with chronic conditions learn to cope or adjust their expectations accordingly; this is known as the disability paradox.13
Is data capture changing?
There is reason to believe that more vigilant recording of conditions accounts for some of the reported rise in chronic disease. In British Columbia, as in many other jurisdictions, physicians receive extra payment for caring for patients with multiple conditions, and at least a portion of hospital funding reflects the complexity of cases. These incentives encourage more comprehensive recording of diagnoses. The largest incentive in British Columbia, for complex care, was introduced in 2006-07, coinciding with the inflection in the lines in fig 1⇑.
Policies and incentives have been shown to affect data recording in other settings. In addition, there is also evidence of greater referral to specialists,14 which increases the number of physician visits and thus the opportunity to record more formal diagnoses of discrete conditions. These phenomena suggest that observed trends result at least partly from more extensive (which may or may not mean more accurate) accounting of conditions in utilisation (administrative) data. This amounts to an apparent increase in diagnosis with no real change in underlying morbidity.
Are we increasing diagnosis and redefining disease?
At the same time, there are widespread efforts to screen populations and catch disease at earlier stages. If this were a main driver of increased disease, we would expect to see higher age specific rates—or at least faster increases in these rates—over time among younger age groups. For example, suppose patients who in 1990 had chronic obstructive pulmonary disease diagnosed at age 75 would have been similarly diagnosed at age 70 in 2000. We would observe an increase in prevalence in the 70-74 cohort in 2000 but not the 75-79 cohort. Instead, figs 1 and 2⇓ show, if anything, more rapid increases in diagnosis at older ages.
A more likely candidate is redefining what constitutes illness, and here there is considerable supporting evidence. Bone density testing has created a vast pool of new patients with osteoporosis. Lowering of clinical thresholds for hypertension and diabetes, among others, has led to substantial increases in the number of people diagnosed.15 16
These trends have prompted a growing concern about overdiagnosis and hypermedicalisation,16 and are part of the motivation for campaigns such as Choosing Wisely and The BMJ’s “Too Much Medicine.”17 18 The moving target is the dividing line between normalcy and disease, and the criteria determining where to draw it.
Where to from here?
The data suggest a growing epidemic of chronic disease, and while British Columbia is a single jurisdiction, the trends reported here are consistent with those seen elsewhere.5 Although there is unlikely to be a single explanation for all of the increase, we argue that changes in data capture and diagnostic practices are far more important than actual changes in health status.
If our conjectures are correct, are patients worse or better off than before? We cannot know for sure unless we assess whether more vigilant recording of diagnoses, lowering diagnostic thresholds, and increasing intensity of services affect long term patient outcomes. More aggressive case finding and intensive management are assumed to be good things. They will certainly lead to more use of healthcare services and higher costs, but evidence suggests that more interventions do not automatically lead to better care and outcomes. The classic American studies found no improvement in disease specific outcomes, physician satisfaction, or patient satisfaction among elderly people in high cost (meaning higher use) regions.19 If anything, higher intensity areas had worse outcomes. This is because nearly all treatments—medical or surgical—have their own risks and unintended consequences. In addition, overzealous case finding may root out conditions that would never progress and cause harm, as is emerging for prostate and breast cancer.20 In healthcare, more is not always better.
This message is underlined by research on the effects of diagnosis on patients. New diagnoses impose monetary and non-monetary costs on patients, including expectations of self management and lifestyle changes, the not inconsiderable time and energy it can take to negotiate care from multiple providers, and the direct cost of new drugs and other therapies.21
So where does this leave us? We are living longer, with more diagnosed chronic conditions, but without necessarily feeling worse (indeed we report that we are feeling better). Healthcare costs are going up because more and more people are diagnosed with conditions; this results in care that in some cases may be doing more harm than good. Encouraging providers to diagnose and treat more conditions is defensible only if there are good prospects for improving lives. The public interest is compromised by both underdiagnosis and overdiagnosis; while underdiagnosis certainly exists (in areas such as mental health), it increasingly seems that overdiagnosis may be the real contemporary pandemic. Differentiating between the two is a critical task for healthcare researchers and practitioners.
WHO defines health as more than the absence of disease and as more than physical attributes. Despite this, our current approach is focused on diagnosis and treatment. If being labelled a patient results in treatment that improves quality of life, the bargain is sound. If it leads to a cascade of anxiety, ineffective interventions, greater expense, and poorer outcomes, it is a disaster.
The rapidly increasing prevalence of diagnosed chronic disease could be due to changes in underlying health; changes in data capture; or changes in healthcare practice
Increasing prevalence seems inconsistent with trends in self reported health
Such large changes in underlying health are unlikely to have occurred within a decade
We must be careful not to conflate diagnosis with illness or morbidity
If changes in healthcare practice are driving the increase, we must be sure they improve outcomes for patients
We thank Sandra Peterson for data analysis, Dawn Mooney for production of figures, and Megan Engelhardt for help in manuscript preparation. Funding for this work came through an operating grant from the Canadian Institutes of Health Research and it was approved by the University of British Columbia behavioural research ethics board.
Competing interests: We have read and understood BMJ policy on declaration of interests and have no relevant interests to declare.
Contributors and sources: Population Data British Columbia provided the data for this analysis. All authors made substantial contributions to the conception or design of the work and the interpretation of data in this viewpoint. KMcG took responsibility for an initial draft of the paper. All authors revised it critically for important intellectual content. All authors provided final approval of the version to be published. All authors agree to be accountable for all aspects of the work.
Provenance and peer review: Not commissioned; externally peer reviewed.