Clinical trials performed exclusively in elderly persons were collected from the interventional trials registered in the ClinicalTrials.gov registry. Based in the United States, this is a publicly accessible, web-based registry that is maintained by the National Library of Congress on behalf of the National Institutes of Health. It represents the most comprehensive trial registry with more than 190,000 trials from 189 countries.[14] Prospective trial registration has become standard practice due to a number of policies and federal legislations that mandate registration of trials.[15–17] As a result, the registry is well-suited to the evaluation of clinical research activity and has been used to assess various aspects of trial design, selective reporting and publication of trials, globalization of the research enterprise, and the scope of study for specific drugs and interventions. [18–23]

We selected all interventional trials with a start date on or after January 1^st, 2006 with participant age eligibility criteria indicating that the trial included exclusively elderly persons, defined as patients ≥65 years of age. We further limited our sample to those conducted primarily in high-income countries because disease burden and the age distributions of patients varies substantially between countries in different income brackets and the proportion of elderly patients is particularly high in high-income countries. All data were downloaded from ClinicalTrials.gov on February 12^th, 2014.

Each trial was manually reviewed and mapped to one of the conditions in the WHO Global Burden of Disease study taxonomy.[24] This classification system includes 155 disease categories. Diseases that do not affect elderly persons in high-income countries were excluded, including several maternal conditions, sepsis and other infectious disorders of the newborn baby, other neonatal disorders, childhood behavioral disorders, sudden infant death syndrome, leprosy, and neglected tropical diseases, leaving 126 diseases for analysis. We excluded trials studying diseases not represented in the WHO classification system (e.g. amyloidosis) or those studying diseases that could not be mapped to a single disease entity (e.g. chronic pain).

Burden of disease data for elderly persons were obtained from the Global Burden of Disease 2010 Study.[25] This study represents a comprehensive, global collaborative project that quantifies the mortality and disability related to diseases, injuries, and risk factors. Data are stratified by age, gender, and country and include a composite measure, disability adjusted life years (DALYs). This metric accounts for both years of life lived in states of disability and years of life lost due to premature death as a result of disease or injury, with one DALY representing one year of healthy life lost. Since the distribution of research activity should reflect both morbidity and mortality, this measure is particularly well suited for the assessment of the alignment between research and disease burden.[26,27]

The Global Burden of Disease 2010 Study provides DALYs for each of the diseases defined in the study. We obtained DALYs for persons ≥65 years of age residing in high-income nations. Furthermore, we identified the disease categories for which ≥80% of the burden of disease in high-income countries is accounted for by persons ≥65 of age as these diseases are most likely to require evidence from trials that focus on elderly patients.

For every trial in our sample, we collected information on the type of intervention under study (e.g. drug, device, biologic, procedure), the outcomes examined (e.g. safety, efficacy, or both), study phase, start year, whether the study was randomized and, if so, the unit of randomization (participant or cluster), the number of participants enrolled (or the estimated number if the study was ongoing), and the length of the study (calculated as the time from study start to primary completion date). The primary funding source for trials is categorized in the registry as industry, government, or other, which includes academic institutions and other non-profit research networks and organizations. We classified each trial as industry or non-industry sponsored.

We identified all trial study site countries and the total number of sites in each trial in order to determine whether a trial was conducted in a high-income country. Countries were assigned to income levels using the World Bank income categorization, which is based on a country’s gross national income per capita.[28] High-income trials were defined as those in which at least half of the study sites were located in a high-income country.

Descriptive analyses were used to characterize the trials according to different design features. Research activity in trials conducted exclusively in the elderly was measured in terms of the number of interventional trials as well as the number of trials that were randomized, that were randomized and studied a drug intervention, and that were funded primarily by industry. We also considered the number of participants enrolled in trials, limiting to those that were randomized at the participant level as opposed to cluster-randomized. For each disease category studied in at least one trial, we calculated the ratio of DALYs per trial in order to compare the allocation of research activity in trials focused exclusively on the elderly across disease categories. The correspondence between research activity and burden of disease was assessed using Spearman’s rank-order correlation. All statistical analyses were performed using SAS software, version 9.3 (SAS institute, Inc., Cary, NC).

There were 877 trials in our cohort that mapped to one of the 126 WHO disease categories pertaining to elderly persons in high-income countries. These trials studied a total of 54 (43%) of the disease categories (Table 3). The 10 most common disease categories evaluated were: ‘Alzheimer’s disease’ (n = 96 trials), ‘other cardiovascular and circulatory diseases’ (n = 78), ‘falls’ (n = 69), ‘other musculoskeletal disorders’ (n = 62), ‘lower respiratory infections (n = 61), ‘other unintentional injuries’ (n = 42), ‘lung cancer’ (n = 36), ‘diabetes’ (n = 33), ‘leukemia’ (n = 33), and ‘other cancers’ (n = 27) (S1 Table). Fewer than half of the disease categories were also studied in RCTs and drug RCTs while primarily industry-funded trials addressed a quarter of all the conditions. In addition to the conditions that were not studied at all, another 19% (24/126) were studied in five or fewer trials.

The burden of disease in the elderly was moderately correlated with the total number of trials (r = 0.53, p<0.001), number of RCTs (r = 0.47, p<0.001), number of drug RCTs (r = 0.40, p = 0.01), and number of industry-funded RCTs (r = 0.47, p = 0.007). There was a weak correlation between burden of disease and number of randomized participants (r = 0.30, p = 0.04).

We found wide variation in the amount of research devoted to different conditions. Among disease categories addressed in at least one trial studying exclusively elderly persons, the number of DALYs per trial ranged from 325 for ‘other nutritional deficiencies’ to 1,677,697 for ‘chronic obstructive pulmonary disease’ (S1 Fig). (‘Other nutritional deficiencies’ represented predominantly malnutrition 15/25 trials] and vitamin D deficiency [9/25 trials]). This represents a 5162-fold range in DALYs per trial. Similarly, when research activity was measured in terms of DALYs per number of randomized participants, there was extreme variation, with the number of DALYs per participant ranging from 1 for ‘varicella’ (which includes varicella zoster) to 14,138 for ‘chronic obstructive pulmonary disease’ (S2 Fig).

There were 5 disease categories for which ≥80% of the disease burden was in persons ≥65 years of age (Table 4). These diseases were addressed in a total of only 117 trials (10.5% of all elderly-only trials) and 87 RCTs (10.4% of all elderly-only RCTs). We found wide variation in the number of trials addressing these conditions with ‘Parkinson’s disease’ not addressed by any trials studying exclusively elderly patients; ‘prostate cancer’, ‘peripheral vascular disease’, and ‘atrial fibrillation’ studied in 11 trials or less; and ‘Alzheimer’s disease and other dementias’ addressed by 96 trials.