Osteoporosis Treatment Gap — A Systematic Literature Review

2.1. Search strategy and information sources

The electronic search strategy was developed using combinations of MeSH terms and free-text keywords related to osteoporosis and treatment gaps. A representative PubMed (9) search strategy was as follows: (“Osteoporosis”[Mesh] OR osteoporosis[tiab]) AND (“treatment gap”[tiab] OR “care gap”[tiab] OR “treatment disparity”[tiab] OR “under-treatment”[tiab] OR “under-diagnosis”[tiab]). Equivalent search strategies were adapted for Scopus (10), Web of Science (11), ScienceDirect (12), and Europe PMC (13). Searches included studies published up to the 7th of April 2025. Full search strategies for all databases are available from the authors upon reasonable request. Grey literature was not systematically searched; however, reference lists of included studies were screened for additional eligible publications. A total of 1457 references were imported to the Systematic Review Accelerator (SRA) platform (14), where duplicates were removed using the Deduplicator Tool (15). Non-English studies were excluded during initial screening due to feasibility constraints.

2.2. Eligibility criteria

Studies were eligible if they met the following criteria:

1. Addressed the osteoporosis treatment gap;

2. Investigated reasons for the treatment gap;

3. Possibly provided an estimate of the treatment gap size.

Adult populations at risk of osteoporosis, with or without prior fragility fractures, were included. “Treatment” was defined as pharmacological therapy for osteoporosis as reported in individual studies. There is no universally accepted definition of the “treatment gap” in osteoporosis. Across studies, it variably refers either to non-treatment following a fragility fracture or to non-treatment among individuals at high fracture risk (based on FRAX, with or without DXA). We therefore did not impose a single definition, but extracted and reported study-specific definitions and interpreted findings within their respective contexts.

2.3. Screening and selection process

Two authors (K.G.R. and L.P.) independently screened titles and abstracts. Discrepancies were resolved through discussion. Full texts of potentially eligible studies were then retrieved and assessed independently for inclusion by the same reviewers. Non-original research articles and non-English publications were excluded. The selection process was recorded in sufficient detail to complete a PRISMA (7) flow diagram (Figure 1).

2.4. Data extraction

A standardised data extraction form was developed and piloted on a subset of included studies. Data were independently extracted by two reviewers (K.G.R. and L.P.) using Microsoft Office 365 Excel (16). Missing data were not sought from the study authors. Given the heterogeneity of included studies, a narrative thematic synthesis was conducted.

2.5. Risk of Bias Assessment

Risk of bias (RoB) was assessed using the Newcastle-Ottawa scale for cohort (17) and cross-sectional studies (18) and the Critical Appraisal Skills Programme (CASP) checklist for qualitative studies (19).

3.1. Study analysis

Of the included studies, 5 were prospective cohort studies (20, 29, 37, 41, 43), 22 were retrospective cohort studies (21,22,23,24,25,26,27,28, 30,31,32,33,34,35,36, 38,39,40, 42, 58, 61–62), 15 were cross-sectional studies (44,45,46,47,48,49,50,51,52,53,54,55,56,57, 60), and 1 was a qualitative study (59). 11 studies were conducted in European countries (27,28,29,30,31, 45,46,47,48, 50, 57), 17 in the United States and Canada (24, 33–34, 36,37,38,39,40,41,42,43, 51,52,53, 58, 60–61), 9 in Asian countries (21,22,23, 25–26, 32, 35, 49, 62), 3 in Australia (20, 56, 59), 2 in northern Africa (44, 54) and 1 study combined data from the US, Canada and India (55) (Figure 2). Most studies were conducted in secondary or tertiary care settings (21,22,23,24, 31–32, 37, 41–42, 47, 53) or in mixed healthcare settings (20, 25,26,27, 29–30, 33,34,35,36, 38,39,40, 43–44, 46, 49, 51–52, 54,55,56,57,58,59, 61–62), while five studies were performed exclusively in primary care settings (28, 45, 48, 50, 60). The majority of studies included both male and female participants (20,21,22,23,24,25,26,27,28, 30,31,32, 35,36,37, 39–40, 42, 44,45,46,47, 50,51,52, 58,59,60, 62). Seven studies focused specifically on postmenopausal women (29, 33–34, 38, 41, 48–49), one included only male patients (43), and one examined patients with cancer of both sexes (61). Several studies were survey-based and targeted healthcare professionals (53,54,55,56,57). The summary data from the 43 included studies were organised by outcomes.

Figure 1:

3.2. Primary outcome

The factors contributing to the treatment gap were categorised (Table 1). Patient-level factors were defined broadly to include both demographic characteristics and clinical or disease-related factors, as these were not consistently distinguished across studies.

3.3. Secondary outcome

The treatment gap estimates are presented in Table 2. Because treatment gap estimates were derived from clinically distinct populations, including post-fracture secondary prevention cohorts and broader osteoporosis/high-risk populations, direct comparison of estimates across these categories should be interpreted with caution.

3.4. Risk of bias assessment

For cohort studies, NOS scores (17) of 7–9 were considered “good”, 5–6 “fair”, and ≤ 4 “poor”. For cross-sectional studies, adapted NOS scores (18) of 8–10 were considered “good”, 6–7 “fair”, and < 6 “poor”. CASP checklist (19) gives a descriptive rather than a numerical grade. Overall, 32 of 43 studies were rated as “good” quality (21,22,23,24,25,26,27,28,29,30, 33, 35,36,37,38,39,40,41,42, 44,45,46,47,48,49,50,51,52, 56,57,58,59, 61), 8 as “fair” (20, 32, 34, 43, 53, 55, 60), and 4 as “poor” (31, 54, 57, 62). The use of non-validated measurement tools, the lack of sample size justification, and inadequate reporting of response rates limited the two cross-sectional studies, which were rated as poor quality (54, 57). The two cohort studies rated as poor quality (31, 62) were affected by issues including non-representative sampling, failure to ensure absence of the outcome at baseline, reliance on self-reported data, absence of a non-exposed cohort, and insufficient adjustment for confounding factors. Risk of bias assessments informed interpretation, with greater emphasis placed on higher-quality studies. Determinants from studies with a higher risk of bias were generally consistent with those from higher-quality studies; however, all findings were interpreted with caution.

Figure 2:

4.1. Treatment gap estimates

This systematic review found reported treatment rates ranging from 38.8% to 94.6% across studies. However, comparisons between studies are limited by substantial methodological heterogeneity. To facilitate interpretation, treatment gap estimates were classified according to whether they were derived from post-fracture secondary prevention populations or broader osteoporosis/high-risk populations. Nevertheless, direct comparisons between these categories should be interpreted with caution due to differences in study populations, eligibility criteria, and treatment gap definitions.

There is no uniform definition or approach to calculating the osteoporosis treatment gap. Some studies assessed treatment following fragility fractures as a form of secondary prevention (20,21,22,23,24,25,26,27,28,29,30,31,32, 36–37, 39,40,41, 43, 47, 56, 58, 61–62), while others included broader high-risk populations, including individuals defined by DXA criteria or FRAX-based risk (29). In several studies, the timing of fracture or treatment initiation was unclear (34–35, 38, 44,45,46, 48, 50,51,52).

Despite these differences, similar treatment gaps were observed across clinically distinct groups. For example, one study reported comparable gaps in postmenopausal women defined by T-score ≤ −2.5 and those defined by prior major osteoporotic fracture (29), suggesting that undertreatment is systemic across both diagnosed osteoporosis and high-risk populations.

Overall, variability in definitions and methodologies limits direct comparison and highlights the need for standardised approaches to defining and reporting the treatment gap.

4.2. Factors contributing to the treatment gap

The identified determinants were grouped into patient-, physician-, and system-level factors to aid interpretation; this categorisation is descriptive and does not imply causal relationships.

4.3. Limitations and strengths

This systematic review has several limitations. First, the review was restricted to English-language publications, which may introduce language bias. This is particularly relevant in a global context, as non-English studies may capture underrepresented healthcare settings and system-level barriers. Consequently, the findings may not fully reflect the global diversity of osteoporosis care and may have limited generalisability. Second, no additional data were sought from the original study authors, which may have limited the completeness of extracted data and the risk of bias assessment. Although the risk of bias was systematically assessed, its impact on the interpretation of findings should be considered. Studies rated as having a higher risk of bias were retained to ensure a comprehensive overview, but were interpreted with caution. Most determinants identified in fair- and poor-quality studies were also reported in studies of good methodological quality, including male sex, ethnicity, comorbidity burden, prior osteoporosis diagnosis or bone mineral density testing, treatment costs, and the lack of routine fracture risk assessment. In contrast, several healthcare professional and organisational factors, including limited familiarity with osteoporosis treatment, knowledge gaps, workload concerns, perceptions regarding responsibility for osteoporosis management, referral and prescription practices, and constraints related to specialist-led treatment plans, were identified primarily in studies with a higher risk of bias (53, 57, 60, 62) and should therefore be interpreted with greater caution. Nevertheless, these factors may represent important hypotheses for future research.

The observational design of included studies, variability in study quality, heterogeneity in outcome definitions, population characteristics and follow-up time limit the strength of inference.

However, the review has notable strengths. It employed a comprehensive search strategy across five major databases, ensuring broad coverage. It includes a diverse, multinational dataset comprising 43 original research studies (20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62) from a wide range of healthcare systems and populations, thereby improving the findings' generalizability. The categorisation of factors into patient, physician, and system-level domains allowed for a structured analysis of the multifactorial nature of osteoporosis under-treatment. RoB was systematically assessed using validated tools (17,18,19), adding methodological rigour.

4.4. Contributions to the field

This review synthesises evidence on gaps in osteoporosis treatment across multiple healthcare systems, providing a comprehensive overview of the determinants of undertreatment. By structuring these into patient-, physician-, and system-level domains, it highlights the multifactorial nature of the problem.

It also demonstrates substantial international variability in post-fracture care, reflecting differences in healthcare organisation, clinical practice, and access to fracture prevention services. Physician-level barriers, such as uncertainty about treatment initiation and inconsistent care pathways, represent actionable targets for intervention. Finally, the review identifies key research gaps, including the optimisation of care for men, improved risk stratification, and the development of scalable post-fracture care models.