When One Illness Overshadows Another: Chronic Disease Patients May Miss Out on Treatment for Other Conditions

Table of Contents

• Key Points
• Why This Research Matters
• How the Study Was Conducted
• Key Finding 1: Diabetes and Estrogen Replacement Therapy
• Key Finding 2: Emphysema and Cholesterol-Lowering Medications
• Key Finding 3: Psychotic Syndromes and Arthritis Treatment
• When Patients Were NOT Overlooked
• What These Results Mean for Patients
• Study Limitations
• Recommendations for Patients and Clinicians
• Frequently Asked Questions
• Source Information

Key Points

• In 1,344,145 Ontario seniors with free drug coverage, chronic disease patients were less likely to get unrelated treatments.
• Diabetes was tied to 60% lower odds of estrogen therapy; emphysema to 31% lower odds of cholesterol-lowering drugs.
• Psychotic syndromes were associated with 41% lower odds of arthritis medication use.
• Breast cancer patients got glaucoma treatment normally, and hypothyroid patients got antibiotics normally.
• Undertreatment was not due to cost or access, since all patients had free medication coverage.

Why This Research Matters

Patients can have several illnesses at the same time. Yet some of those diseases may be neglected when one problem consumes all the attention. "Nature has no mercy," the authors write — one disease provides no immunity against others. In a large population, the coincidental occurrence of two unrelated diseases must happen often.

But human reasoning has blind spots. One common error is the "gambler's fallacy" — the misconception that bad luck tends to be followed by good luck. In medicine, this might create a belief that unusual coincidences are extremely unlikely in any individual patient. This belief can be reinforced by Occam's razor, a scientific principle that urges investigators to use the simplest explanation to explain all observed facts.

Research in nonmedical situations shows that reasoning can be inconsistent even when two problems are clearly separate. For example, a $200 cost seems less expensive when placed next to a $20,000 expenditure than when considered alone. This decreased cost-consciousness helps explain why consumers make imprudent choices, such as purchasing an overpriced radio when buying a reasonably priced automobile.

More generally, secondary problems may receive too little attention when they occur alongside a larger problem. Medical training encourages clinicians to consider a broad differential diagnosis (the full range of possible conditions) in each patient. One classic teaching dictum states: "Even psychiatric patients can have surgical diseases." However, the authors hypothesized that fundamental mistakes in reasoning are not eliminated by formal medical training. Their specific question: Are patients less likely to receive treatment for selected unrelated disorders if they have a chronic medical disease than if they do not?

How the Study Was Conducted

The study took place in Ontario, Canada, in 1995. Ontario was chosen because it is Canada's largest province, home to about one third of the country's population. That year, Ontario had a population of 11,008,400, with 25,624 licensed physicians, 6,201 licensed dentists, and $3.5 billion in Ontario Drug Benefit expenditures.

All residents aged 65 or older were included — a total of 1,344,145 people (56 percent were women, with a mean age of 74). These patients received prescription medications free of charge through the Ontario Drug Benefit program, meaning financial barriers to treatment did not exist. This universal coverage made Canada an ideal setting to test whether chronic disease leads to undertreatment of unrelated conditions.

How Patients with Chronic Diseases Were Identified

Researchers selected three chronic systemic medical diseases that are complicated to manage and identifiable through medication prescriptions:

• Diabetes mellitus (diabetes) — identified by any prescription for insulin
• Pulmonary emphysema (a chronic lung disease that damages the air sacs) — identified by any prescription for ipratropium bromide (an inhaled medication)
• Psychotic syndromes (severe mental health conditions involving hallucinations or delusions) — identified by any prescription for haloperidol (an antipsychotic medication)

These identification methods were imperfect. Some patients with emphysema never receive ipratropium bromide (false negatives), and some without emphysema do receive it (false positives). Both types of errors tend to weaken statistical comparisons and can bias the analysis toward finding no differences — making any observed differences more meaningful.

Choosing the Unrelated Treatments

For each chronic disease, the researchers selected one unrelated treatment that met four criteria:

1. The treatment was not directly related to the cause or management of the chronic disease.
2. The treatment was not indirectly related through a shared underlying risk factor.
3. The treatment was relatively straightforward but possibly less important than the chronic disease therapy.
4. The treatment was not trivial and would require follow-up.

The three pairings chosen in advance were:

• Estrogen-replacement therapy (hormone treatment for menopause symptoms) for patients with diabetes
• Lipid-lowering medications (cholesterol-lowering drugs, including statins) for patients with emphysema
• Medical arthritis treatment (medications for joint inflammation) for patients with psychotic syndromes

Comprehensive Medication Lists

The researchers used an exhaustive approach to identify these treatments. Estrogen-replacement therapy could be oral or transdermal (skin patch) and included conjugated estrogens, esterified estrogens, 17-beta-estradiol, and combinations containing ethinyl estradiol.

Lipid-lowering medications included cholestyramine, colestipol, clofibrate, bezafibrate, fenofibrate, gemfibrozil, lovastatin, pravastatin, simvastatin, and fluvastatin.

Medical arthritis treatment included auranofin, aurothioglucose, gold sodium thiomalate, diclofenac, etodolac, fenoprofen, floctafenine, flurbiprofen, hydroxychloroquine, ibuprofen, indomethacin, ketoprofen, ketorolac, naproxen, piroxicam, sulindac, tenoxicam, tiaprofenic acid, and tolmetin. This category included both NSAIDs (nonsteroidal anti-inflammatory drugs, such as ibuprofen) and DMARDs (disease-modifying antirheumatic drugs, such as gold compounds and hydroxychloroquine).

Supplementary Analyses and Statistical Methods

Two additional analyses tested situations where researchers expected no difference. First, they identified patients with a history of breast cancer (by prescriptions for tamoxifen) and checked whether they received glaucoma treatment (eye drops including betaxolol, carbachol, dipivefrin, echothiophate, levobunolol, pilocarpine, and timolol). Second, they identified patients with hypothyroidism (underactive thyroid, by prescriptions for thyroxine) and checked whether they received antibiotic treatment for acute infections. Both pairings were selected in advance, and the comparisons were prespecified.

All comparisons were two-tailed and expressed as odds ratios (a measure of how much less or more likely treatment was). The researchers used the chi-square test to compare proportions and logistic regression (a statistical method) to adjust for imbalances in age and sex. The estrogen analysis excluded all men. To check reliability, all 1995 comparisons were replicated using separate 1995 data from 1990, coded and analyzed by the same methods.

Patient confidentiality was protected with unique encrypted numbers that allowed data linkage while preserving anonymity. Medications were coded by Drug Identification Number from claims filed electronically by pharmacies.

Key Finding 1: Diabetes and Estrogen Replacement Therapy

The study included 30,669 patients with diabetes mellitus (55 percent women, mean age 74). These patients were significantly less likely to receive estrogen-replacement therapy than other patients in the study.

The numbers tell the story clearly: only 2.4 percent of women with diabetes received estrogen replacement, compared with 5.9 percent of women without diabetes (P<0.001). In absolute terms, that is about 24 of every 1,000 women with diabetes receiving treatment, versus 59 of every 1,000 women without diabetes.

In statistical terms, diabetes was associated with a 60 percent reduction in the odds of receiving estrogen treatment (odds ratio, 0.40; 95 percent confidence interval, 0.37 to 0.43). A confidence interval is a range that likely contains the true value; because this range stays far below 1.0, the finding is robust. The 60 percent reduction persisted after adjusting for age and after restricting the analysis to women (99.5 percent of those who received estrogen prescriptions were women). The reduction appeared for both oral and transdermal preparations, and analysis of 1990 data showed a similar 61 percent reduction (95 percent confidence interval, 55 to 66 percent).

Key Finding 2: Emphysema and Cholesterol-Lowering Medications

Overall, 56,779 patients had pulmonary emphysema (46 percent women, mean age 76). These patients were less likely to receive lipid-lowering medications than other patients.

Only 6.3 percent of emphysema patients received cholesterol-lowering drugs, versus 8.7 percent of other patients (P<0.001). In absolute terms, that is about 63 of every 1,000 emphysema patients, compared with 87 of every 1,000 patients without emphysema.

Emphysema was associated with a 31 percent reduction in the odds of receiving lipid-lowering treatment (odds ratio, 0.69; 95 percent confidence interval, 0.67 to 0.72; P<0.001). The reduction persisted after adjusting for both age and sex (a 23 percent reduction), was found for both statin and non-statin agents (32 percent and 21 percent reductions, respectively), and was also present in 1990 (35 percent reduction).

Key Finding 3: Psychotic Syndromes and Arthritis Treatment

Similarly, 17,336 patients had psychotic syndromes (64 percent women, mean age 80). These patients were significantly less likely to receive medical treatment for arthritis.

Just 18 percent of patients with psychotic syndromes received arthritis medication, compared with 27 percent of other patients (P<0.001). In absolute terms, that is 18 of every 100 patients versus 27 of every 100 patients — a substantial gap for a painful, treatable condition.

Psychotic syndromes were associated with a 41 percent reduction in the odds of arthritis treatment (odds ratio, 0.59; 95 percent confidence interval, 0.57 to 0.62; P<0.001). The reduction persisted after adjusting for age and sex (38 percent), appeared for both NSAIDs (41 percent reduction) and DMARDs (55 percent reduction), and was present in the 1990 data as well (40 percent reduction).

When Patients Were NOT Overlooked

The relative neglect of unrelated disorders was not a universal finding. Two situations showed that doctors did treat separate conditions normally.

The 11,094 patients with breast cancer were just as likely to receive glaucoma treatment as other patients (5.1 percent versus 4.8 percent, P>0.20 — not a statistically significant difference). The 140,460 patients with hypothyroidism were slightly more likely to receive treatment for acute infections than other patients (56 percent versus 54 percent, P<0.001). Apparently, these unrelated diseases were distinct enough that patients received treatment regardless of their chronic condition.

Cross-Comparisons Reveal the Full Picture

The table below shows all nine cross-comparisons from the study. Each comparison used data from all 1,344,145 patients. For example, the first entry compares 30,669 women with diabetes against the rest of the population.

Different pairings yielded different results — and one positive association. Patients with diabetes were actually more likely to receive lipid-lowering therapy than others (11.4 percent versus 8.5 percent). This makes medical sense: hyperglycemia (high blood sugar) is associated with dyslipidemia (abnormal blood fats). However, the observed 38 percent relative increase was smaller than the roughly 100 percent increase expected based on some epidemiology studies. In other words, even the "correct" association was smaller than it should have been.

Patients with emphysema had a 29 percent relative reduction in estrogen treatment, even though smoking is a known risk factor for emphysema, heart disease, and osteoporosis (thinning bones) — meaning this group might have deserved more estrogen, not less. Emphysema patients had only a 7 percent relative reduction in arthritis treatment, possibly because lung symptoms feel distinct from joint symptoms. Patients with psychotic syndromes, however, were consistently unlikely to receive any of the three unrelated treatments.

What These Results Mean for Patients

The most important finding is the inverse correlation (an opposite relationship) between having a chronic disease and the likelihood of receiving treatment for an unrelated disorder. In no case did the chronic disease justify withholding an effective medical treatment.

The results are compatible with the theory that one disease provides "protection" against other diseases — but that theory is unlikely to be correct, given medical pathophysiology and shared underlying risk factors. Instead, the findings suggest a shortfall in health care: unrelated disorders are relatively neglected in patients with chronic medical diseases.

The chosen unrelated treatments had important implications for each chronic disease:

• Diabetes: Patients with diabetes are at increased risk for atherosclerosis (hardening of the arteries) and may be particularly likely to benefit from estrogen-replacement therapy.
• Emphysema: The reserve capacity of patients with emphysema is seriously compromised. They may be unable to tolerate even a small cardiovascular event (such as a heart attack or stroke), making cholesterol treatment especially valuable.
• Psychotic syndromes: These patients are often sensitive to discomfort, and joint pain might theoretically worsen their mental status. Treating arthritis could protect their mental health as well as their physical comfort.

In all three examples, inadvertent undertreatment may have real consequences. These examples resemble other reported cases of mistakes in caring for patients who have more than one illness.

The study found, however, that undertreatment cannot be blamed on cost or paperwork. The results cannot be attributed to a general tendency toward prescribing multiple medications for the elderly, since that would work against finding any negative associations. Nor can they be blamed on barriers to medical access, insurance status, or ability to pay — all patients had free drug coverage. Fraud (such as multiple people sharing one insurance card) also does not explain the findings.

Study Limitations

The authors acknowledge several important limitations, of which three merit emphasis.

First, this was not a randomized trial. It is impossible to randomly assign patients to have or not have a chronic disease. Subtle confounding (hidden factors that distort results) could have contributed to — and possibly justified — the observed differences. However, imbalances related to age, sex, insurance status or carrier, ability to pay, or random chance would not explain the findings.

Second, optimal rates of these treatments are controversial. In theory, the findings could be explained by overtreatment of patients who do not have chronic diseases. If true, that would represent a different — and potentially more common — failure in medical decision making.

Third, the underlying mechanism remains unknown. Future research must determine whether the second disease is simply not detected in the presence of the first, or whether it is detected but not treated.

The observed results might arise from several sources. Patients with chronic diseases may feel exhausted and reluctant to accept multiple interventions. Clinicians are often busy and may strive to keep care simple — particularly when they do not have extra time for patients with more complicated conditions. A chronic disease, especially chronic psychosis, may limit communication between patient and doctor. Even universal insurance coverage could contribute if the implicit goal of equity is achieved by "doing something for all but a lot for none."

Recommendations for Patients and Clinicians

It is important to note that unrelated treatments are not always indicated for patients with chronic diseases. The authors offer several cautions:

• Chronic diseases are sometimes associated with reduced life expectancy, making long-term preventive therapy unrewarding.
• Adding supplementary medications increases the risk of unwanted drug interactions and potential adverse events.
• Prescribing additional medications might alter a patient's compliance with essential medications and indirectly cause harm.
• Time constraints, communication problems, patient preferences, and the priorities of specialists can make it difficult to address more than one problem effectively in a single visit.
• It is often sensible to postpone minor treatments until major problems are resolved.

Despite these cautions, the findings highlight a vital role for clinicians who can provide a comprehensive approach to patient care. Contrary to popular opinion, such comprehensive care was not fully evident in the mid-1990s even under the Canadian system of universal health insurance.

For patients, this study offers actionable advice:

1. Keep an updated list of all your medical conditions and bring it to every appointment.
2. Mention new symptoms even if they seem unrelated to your main chronic condition.
3. Ask your doctor directly about preventive treatments — such as cholesterol or bone health medications — if you have a chronic disease.
4. Do not assume a new ache or symptom is "just part" of your existing illness. Ask for an evaluation.

Primary care physicians may be well suited to this coordinating role, since the diversity of their practices makes them responsive to diverse issues. Yet specific training may be required to overcome fundamental pitfalls in reasoning. Alternatively, a health care system might develop programs that ensure comprehensive care and avoid focusing solely on well-defined single problems, while minimizing failures of individual judgment. Although chronic diseases are frequent and unrelated disorders are common, the authors conclude, the inevitable coincidences are relatively neglected in clinical practice. Patients and doctors alike should remember: one disease does not protect against another, and treating the whole patient means looking beyond the most obvious diagnosis.

Frequently Asked Questions

What did this Canadian study find about treating unrelated conditions in chronic disease patients?

In a study of 1,344,145 Ontario seniors with free drug coverage, those with diabetes, emphysema, or psychotic syndromes were significantly less likely to receive treatment for certain unrelated conditions. For example, women with diabetes had lower estrogen therapy use, and emphysema patients had lower cholesterol medication use. This suggests one illness may overshadow other health needs.

Why were Ontario seniors chosen for this study?

Ontario was Canada's largest province and provided free prescription medications to all residents aged 65 or older through the Ontario Drug Benefit program. This eliminated financial barriers, allowing researchers to test whether chronic disease leads to undertreatment of unrelated conditions without cost being a factor. The study included all 1,344,145 eligible seniors.

Which specific chronic diseases and unrelated treatments were studied?

Researchers paired three chronic diseases with unrelated treatments: diabetes with estrogen-replacement therapy, emphysema with lipid-lowering (cholesterol) medications, and psychotic syndromes with arthritis medications. These pairings were chosen in advance because the treatments were unrelated to the chronic disease and not trivial.

Were there any situations where doctors did treat unrelated conditions normally?

Yes. Patients with breast cancer were just as likely to receive glaucoma treatment as others, and patients with hypothyroidism were slightly more likely to receive antibiotics for acute infections. These comparisons were included as controls and suggest that undertreatment was not universal but specific to certain chronic disease and treatment pairings.

What should patients with chronic diseases do based on this study?

Patients should keep an updated list of all medical conditions and bring it to every appointment, mention new symptoms even if seemingly unrelated, and ask directly about preventive treatments. The study advises not assuming a new symptom is just part of an existing illness and requesting evaluation. Discussing all health concerns with your doctor is important.

I have a chronic disease like diabetes, emphysema, or a psychotic syndrome — could a second opinion help make sure I am not missing treatment for an unrelated condition?

In a large Ontario study of older adults with free prescription drug coverage, patients with diabetes, emphysema, or psychotic syndromes were significantly less likely than others to receive treatment for unrelated conditions. For example, diabetes was associated with 60% lower odds of estrogen therapy in women, and psychotic syndromes were associated with 41% lower odds of arthritis treatment. Chronic disease can unintentionally overshadow other health needs. Patients and clinicians should remember that one disease does not protect against another. Diagnostic Detectives Network provides independent expert second opinions.

Source Information

Original article title: The Treatment of Unrelated Disorders in Patients with Chronic Medical Diseases

• Authors: Donald A. Redelmeier, M.D., Siew H. Tan, M.A., and Gillian L. Booth, M.D.
• Journal: The New England Journal of Medicine, 1998; Volume 338, pages 1516-1520 (May 21, 1998 issue)
• Institution: Department of Medicine, University of Toronto; Clinical Epidemiology and Health Care Research Program, Sunnybrook Health Science Centre; and the Institute for Clinical Evaluative Sciences in Ontario, Toronto, Canada

This patient-friendly article is based on peer-reviewed research. It is intended for educational purposes and does not replace individualized medical advice from your healthcare provider.