Osteogenesis imperfecta (OI) is a heterogeneous group of inherited connective tissue diseases defined clinically by excessive skeletal fragility and recurrent fracture [2]. With an estimated prevalence of 8 cases per 100,000 persons, perhaps 30,000 live with OI in the United States [2]. The Sillence classification for OI, first described in 1979, includes four types (I-mild, II-neonatal lethal, III-severe, progressively deforming, and IV moderately severe), and is almost exclusively based on physician perception of disease severity [3].

Most OI cases are caused by mutations in the type I collagen genes and are the result of autosomal dominant inheritance. During the past several years, the spectrum of OI has expanded to include rare, recessively inherited forms resulting from abnormalities in the post-translational modification of type I collagen, as well as abnormalities in regulatory proteins involved in bone matrix homeostasis, including those of the Wny-signaling pathway [4, 5]. These non-classical forms of OI can range from moderate severity to neonatal lethality in presentation.

While the cardinal clinical manifestation of OI is fracture [2], deficiencies in collagen-rich tissues other than bone can lead to abnormal dentition [6], joint laxity [7, 8], hearing loss [9], ocular disease [10], pulmonary [2], and vascular and valvular heart disease [11]. Disease severity ranges from perinatal death to minimally symptomatic forms that escape detection into adulthood [2]. Heightened diagnostic awareness and improved treatments, particularly in severe forms, has increased the number of adults living with OI [12]. While adolescents and young adults with OI may be medically sophisticated, their transition to a new adult health care provider and system of care can be daunting. Adult health care providers and systems have a paucity of medical evidence to guide them and likely little first-hand experience with this condition [13]. Similarly, physically accessible, high quality, informed health care may not be readily available to many [14] and some adults with OI feel that the strong association of OI with “brittle bones” diverts attention away from other important adult health care concerns.

Limited research exists regarding the health issues of adults with OI [7, 8]. Although adults with OI report a high level of life satisfaction, any conclusions drawn from available reports are constrained by the small number of subjects, the focus on milder forms of OI, and the absence of disease specific quality of life (QoL) measures. Thus, adults with OI, their families and their health care providers cannot formulate expectations of their future health-related QoL, nor anticipate their long-term health care needs. Moreover, despite the potential for mortality and significant morbidity suffered by OI patients, all measures of OI outcomes have been developed by experts without input from patients and rely on clinician-based assessments rather than actual patient reports.

Directly engaging the OI community in patient-centered outcomes (PCO) research could mitigate some of these challenges [1, 15]. To address concerns raised by the adult OI community and to begin closing the knowledge gap, the Osteogenesis Imperfecta Foundation (OIF) assembled a committee composed of adults with OI (including a practicing physician), orthopedic and medical specialists with extensive experience in the care of persons with OI, OIF representatives, research scientists, and statisticians and created the Osteogenesis Imperfecta Adult Natural History Initiative (OI-ANHI) in 2010. The OI-ANHI tested the feasibility of using web-based, PCO methods for rare disease research by creating an on-line health survey to (1) define health care concerns and perceptions of adults with OI; (2) identify health-care related issues that may have been previously missed or under-valued by adults with OI and their medical providers; and (3) compare QoL responses by adults with OI with those of benchmark populations without OI. This paper reports the initial results of the OI-ANHI survey.

Based on adult focus group meetings conducted by OI-ANHI committee members at the 2010 bi-annual OIF meeting for persons with OI and their families, a web-based survey was designed and implemented to capture the health status, concerns, needs, and priorities of adults with OI. The complete OI Adult Natural History Initiative survey is available as Additional file 1. The survey was presented in the following sequence: (1) collection of general anthropomorphic, demographic, and medical information; (2) an organ-system-based, medical “review of systems” (ROS) followed by a 5-point Likert ranking to assess the impact of each organ-system on current and expected future QoL; (3) multiple instruments from the Patient-Reported Outcomes Measurement Information System (PROMIS®) to evaluate the level of health and functioning for core physical, mental, and social health constructs that are normed to the US general population [16].

Anthropometric data included height and weight. Demographics included year of birth, gender, race/ethnicity, education level, and current country of residence. Medical information addressed medical diagnoses, testing and interventions, and various lifestyle factors. Data collected include OI Sillence Classification, additional medical diagnoses, primary means of mobility, list of health care providers, surgical health interventions, and level of physical activity among other variables.

We queried severity of OI in three different ways: first, respondents provided their understanding of their Sillence type; secondly, they reported whether they perceived their disease to be mild, moderate or severe; height provided an objective measure that has been shown to correlate with disease severity [17].

The second portion of the survey was a medical review of systems (ROS), developed as a collaborative effort by the group of authors, which included physicians, persons with OI, and their advocates, yielding a comprehensive, textbook-quality medical ROS and potentially identifying previously underemphasized or unidentified health concerns. (Additional file 1) We assessed respondents’ physical health within fifteen organ systems: neurologic, urinary, musculoskeletal, auditory, cardiovascular, skin, vision, dental, pulmonary, gastrointestinal, endocrine, hematologic, oral, obstetric/gynecologic (females only), and male sexual function (males only). We then used a tri-part format to assess the impact of each system on the respondent’s QoL. The first question asked whether or not the respondent had any of the concerns listed specific to the system (e.g. for the skin, rashes and pressure sores; for the pulmonary system, shortness of breath, pneumonia, etc.). The second and third questions for each organ system queried respondents via a 5-point, Likert rating scale as to how much issues within the relevant organ system currently impacted their QoL and how much they thought problems with each system would impact their quality of life in the future.

The third and final portion of the survey used selected PROMIS® item banks to evaluate general physical health, pain, fatigue, physical function, sleep, depression, anxiety, sexual satisfaction, and social function. Point values for PROMIS® construct responses were calibrated to T-scores generated from a representative U.S. population and reported on a 100-unit scale where the mean is set at 50 units, with 10 units representing one standard deviation (SD). This data transformation follows the guidelines set forth in PROMIS® methodology [18]. Means and 95 % confidence intervals for the OI population were calculated based on individual T-scores for a given concept. The minimum important difference (MID) in PROMIS® responses thought to be clinically relevant has been estimated as 0.3–0.5 SD for fatigue, anxiety and depression; 0.4–0.6 SD for pain; and 0.2 for physical function [19, 20].

Participants accessed the survey through a web-based portal and were recruited with two OIF postal mailings, a notice posted on the OIF web-site, the OIF newsletter Breakthrough, spoken and written notification at national and regional clinical and scientific meetings, social networks (Facebook, Twitter, OI forums and chat-rooms), and word of mouth. Children’s National Health System’s Institutional Review Board exempted the survey from review, as it was anonymous. Prior to participation, respondents read the study’s purpose, gained study contact information, learned of the right to withdraw, and then chose either to participate or not participate. The survey instrument was available from 10/01/2011 to 12/31/2011 and designed so that respondents could complete the survey progressively. Those without computer access or facility were offered a paper version, and a data entry specialist masked to subject identity uploaded completed results. Paper surveys were accepted until midnight 01/02/2012. The survey required at least one hour, but more than likely, two hours for completion. Only respondents who answered at least one question in the final PROMIS® item set (Sleep Disturbance) were considered to have completed the entire OI-ANHI survey.

Body Mass Index (BMI) was calculated as kg/m² and classified according to the World Health Organization (WHO) weight classification criteria [21]. Height Z-scores were calculated from the National Health and Nutrition Examination Survey (NHANES III) dataset [22]. Height Z-scores indicate the number of SD a reported height varies from a US adult, gender matched mean. To compare the PROMIS® scores of the generalized population with those of our sample, we segregated PROMIS® T-scores by both OI self-reported severity grouping (mild, moderate, severe) and by quartile of height Z-score. For continuous variables, a t-test was used to calculate the p-value; for dichotomous variables, a chi-square test was used. For variables reported on a Likert scale, a generalized linear model was utilized (SAS proc glm). Generalized linear models were also used to compare differences among OI groups. All analyses were performed using SAS version 9.2 (SAS institute, Cary NC).

Of 1,183 individuals who initiated the survey, 959 (82 %) successfully completed it, constituting the study sample. Demographically, respondents were female (71 %), middle aged (mean 45.1 ± 15.0 years, range 18–89 years), Caucasian (91 %) and lived in North America (89 %). Figure 1 is a graphical depiction of age distribution. Age did not vary on the basis of self-reported severity grouping (Fig. 2). Highest educational achievement exceeded secondary school for 82 %. In regard to anthropometric data, average height was 57.3 ± 9.2 inches, weight 135.9 ± 44.6 pounds, and BMI 29.4 ± 9.3.

In response to the OI-specific medical questions, the majority walked unaided (61 %), but 16 % required a cane, crutches or walker, and nearly one-quarter reported regular wheelchair use. The majority had received their OI diagnosis from a doctor (59 %). Respondents considered their OI to be mild (56 %), moderate (35 %) or severe (9 %) and reported Sillence classifications of Type I (39 %), Type III (13 %), and Type IV (12 %). Fewer than 5 % described other Sillence types and 31 % reported their type to be unknown. Table 1 shows the relationship between self-reported disease severity and self-reported Sillence type. As self-reported disease severity worsened, height Z-score decreased (β = -0.13, p < 0.0001, Fig. 3).

		Self-reported disease severity
		n (% of OI type)
OI type	n (%)	Mild	Moderate	Severe
I	369 (39)	308 (83)	59 (16)	2 (<1)
II	26 (3)	12 (46)	8 (31)	6 (23)
III	125 (13)	7 (1)	67 (20)	51(57)
IV	110 (12)	45 (41)	55 (50)	10 (9)
V	14 (1)	2 (14)	11 (79)	1 (7)
VI	2 (<1)	–	1 (50)	1 (50)
Bruck’s syndrome	3 (<1)	–	3 (100)	–
Unknown	295 (31)	154 (52)	123 (42)	18 (6)
Total	944 (100)	528 (56)	327 (35)	89 (9)

Missing n = 15

A large proportion of participants (65 %) consumed vitamins or dietary supplements, followed by pain (40 %) and blood pressure medications (32 %). Medical providers reported to be actively involved in respondents’ medical care included a “general medical doctor” (92 %), orthopedist (49 %), gynecologist (40 % of women), and endocrinologist (21 %). Cardiologists, dermatologists and gastroenterologists were involved in 10 % of respondents’ care, but few other medical specialists had ever been engaged. Other providers were dentists (68 %) and ophthalmologists (30 %). Both overall and when separated by self-reported severity grouping, respondents saw a median number of 4 different care providers. The overall range was 0 to 14 (Fig. 4). Number of health care providers seen did not vary on the basis of self-reported OI severity grouping. Notably, just over half of participants reported no health interventions (52 %), while 35 % had received rodding surgery. Other than the two aforementioned responses, respondents indicated a scant number of other health interventions. Many of the respondents had received medical testing in the prior year with 86 % reporting a blood pressure reading, 71 % a vision exam, 66 % a blood test for cholesterol, and 34 % a DXA scan. Respondents reported very few cancer diagnoses, with the majority reporting none (90 %) followed by skin cancer (4 %).

Respondents acknowledged few general health concerns, with “none of these” (52 %) as the most endorsed response followed by changes in sleep patterns (27 %) and mood swings (20 %). Almost half of respondents (49 %) engaged in walking as a regular physical activity although approximately one-third (34 %) reported that they do not exercise regularly. There was a significant relationship between engagement in physical activity and severity grouping (Fig. 5). The most popular exercise venue was the home (51 %), and respondents oftentimes developed their own exercise programs (40 %).

Only 2 % of respondents reported having no medical provider and 4 % considered urgent care or emergency room providers to be their primary source of health care. Ninety-one percent of respondents had health care insurance. Many adults with OI turn to their medical providers for health-related information (53 %); however, more rely on the Internet (63 %) and the OIF (71 %) for that information. Indeed, only 32 % indicated “quite a bit” or “very much” confidence in their primary care provider’s medical management of their OI. Approximately one-third of respondents reported no regular dental care.

The organ systems most often reported to affect current QoL are urinary tract (97 %), musculoskeletal (95 %), vision (82 %), auditory (75 %), dental (74 %), skin (68 %), gastrointestinal (65 %), and neurological (64 %). Using a Likert scale, Fig. 6 shows the impact of specific organ system concerns on current and anticipated future QoL for all respondents (Fig. 6 left) and for those already experiencing problems within that organ system (Fig. 6 right). Disproportionate concern was reported for future pulmonary, cardiologic and vision issues by all respondents, and for endocrine, pulmonary and cardiovascular issues in those respondents already experiencing issues in these domains.

Figure 7 shows the mean PROMIS® T-scores of selected items by self-reported OI severity compared with the PROMIS® population norm. General physical health status was significantly lower than that of the general population and the magnitude of difference (0.4–1.0 SD) exceeded the MID established for several PROMIS® domains [19, 20]. General physical health among the OI cohort was lower in those reporting greatest disease severity but the magnitude of this difference was small. Results also revealed small but statistically significant differences between adults with OI and the normative population for a number of the PROMIS® scales; adults with OI reported greater levels of anxiety (p < .0001) and depression (p < .0001), as well as lower levels of general mental health (p < .0002). However, the differences in means between adults with OI and the normative population for these three scales did not qualify for clinical relevance according to the established MIDs [19, 20]. Segregation of PROMIS® T-scores by quartile of height Z-score (Fig. 8) demonstrates results similar to Fig. 7 and shows that decreased height does not consistently correlate with less desirable PROMIS® scores within our adult OI sample for certain scales, including access to help and satisfaction with social roles.

Remarkably, adults with OI report lower pain intensity than the general PROMIS® population. While adults with OI indicate that pain interferes with their functional state significantly more than it does for the general PROMIS® population, this difference appears to be less than the MID and results did not differ by self-reported disease severity.

Only those adults with severe OI appeared to report notably (-0.56 SD) lower satisfaction with sexual activity. Statistically significant differences in T-scores for responses to “sleep disturbance”, “satisfaction with social roles” “fatigue”, “access to help”, and “ability to participate in social activities” were found between the OI cohort and the general PROMIS® population but these differences appear to be less than a MID.

Delineating and meeting the health needs of adult survivors of childhood-onset conditions is a growing quality-of-care issue. Based on estimates of those living with OI in the US ranging from 25,000 [39] to 50,000 [40], approximately 3 % of the total US OI population completed the OI ANHI survey. This number is especially impressive given the fact that adults with OI are a rare, geographically dispersed population and that we recruited respondents in an incentive-free manner. In doing so, we have demonstrated how persons with rare (and not so rare) conditions can collectively address some of their important health related concerns as a virtual community.

The OI-ANHI survey engaged an adult OI community to identify issues of specific health-related QoL concern and to focus future research efforts designed to improve quality of care. Our results revealed that while adults with OI exhibit similarities with the normative population, there are important differences, including a higher prevalence of musculoskeletal and auditory problems and specific QoL concerns regarding the musculoskeletal, auditory, pulmonary, and endocrine systems. Our future research will focus on expanding the range of health care topics pursued in OI research and on incorporating patients’ QoL concerns into effectiveness evaluations of diagnostic and treatment strategies. Finally, we will strive to develop methodologies that allow us to be more inclusive and engage the broader OI community in our future research studies.

The data supporting the results of this article are housed in the Osteogenesis Imperfecta Foundation.