Interventions:
The interventions were generally well described and appeared appropriate. Decision thresholds were not reported for FRAX risk assessment but there were numerous references to further FRAX risk assessment tool information.
Effectiveness/benefits:
Data used to generate model results were from large high-quality representative UK datasets. However, the analysis conducted using these data had several flaws. The authors' comparison did not report sensitivity and specificity of the strategies. The total number of expected hip fractures was reported but true and false positives and negatives could have been reported and were not. The number of averted fractures was not reported. NOGG identified approximately half as many women at ages 75 and up and the average BMD t-score for those identified was lower; this indicated that on average NOGG identified higher-risk women.
Costs:
The study choice of cost data was from an appropriate and recent UK source but costs included in the model were insufficient. Only costs of the interventions were estimated; costs of hip fractures were not estimated. This meant that a key cost element in assessing the value of the different interventions was not considered. Because NOGG identified fewer hip fractures in elderly individuals, there would likely be additional costs for the hip fractures that would be prevented under RCP but not under NOGG. The authors acknowledged that NOGG identification of fractures could improve in the elderly but they did not quantify what costs resulted from lower specificity and this hid potential costs associated with the case-finding strategy.
Analysis and results:
The results did not analyse the effect that detecting fewer fractures could have in elderly women. The authors acknowledged that the analysis was limited by only evaluating hip fractures, which they indicated make up 14% of osteoporosis fractures.
There was no indication that the model followed patients over time so the benefits and cost savings from preventing hip fractures were not captured by the model.
Concluding remarks:
The study did not appear to set out to conduct a full cost-effectiveness analysis. As such, insufficient costs and benefits were accounted for or reported in the analysis to fully evaluate cost-effectiveness.