OBJECTIVES: Inadequate reporting of outcomes is a major problem while performing meta-analysis. The objective of this study is to estimate the nature and scale of inadequate reporting of dichotomous outcomes. METHODS: Reporting quality of dichotomous outcomes (efficacy and safety) was analysed across six systematic reviews conducted in four disease areas namely oncology (non-small cell lung cancer, colorectal cancer and ovarian cancer), overactive bladder, multiple sclerosis and rheumatoid arthritis. Reporting quality was considered inadequate when either number of patients analysed (N) was not reported or percentage was reported improperly. Reporting of percentage was considered as improper when the number of patients with outcome (n) could not be calculated from the given percentage. This may be the case when more than one value corresponds to same reported percentage (example 5% of 498 could have values between 23 and 27). Analyses were conducted using STATA 9.2. RESULTS: In total, 6408 reported outcomes were included in the analysis. Inadequate reporting of dichotomous outcomes was frequent and observed for 691 (10.78%) outcomes. The estimate of inadequate reporting varied across selected reviews and ranged from 7.36% to 17.41% in overactive bladder and multiple sclerosis, respectively. Estimate of inadequate reporting was similar for the safety and efficacy outcomes (10.53% and 11.19%, respectively). Improper reporting of percentage contributed for approximately two-third (65.56%) of the inadequately reported outcomes whereas for the remaining outcomes, N was missing. CONCLUSIONS: Inadequate reporting of dichotomous outcome was frequently observed among the selected reviews. It was observed that reporting of percentages in published reports are often imperfect (due to rounding of values). This leads to loss of data available for meta-analyses as number of patients with outcome could not be calculated accurately. Our results are indicative of outcome reporting bias which needs to be investigated further.