TABLE 3.

Relative size of impact of hypoglycaemia in included studies

Authors (year) a	Reference	Analysis	Quantified impact on measures of quality of life
Ali et al. (2012)	35	Linear regression	No; 1+ severe hypoglycaemic events did not lead to changes in mental or physical health; compared to those reporting no events, those reporting severe events experienced a decline in SF‐36 Mental Component scores, B = –2.14, and a rise in SF‐36 Physical Component scores, B = 1.05, though none of these impacts were significant, ps > 0.05
Briggs et al. (2017)	20	Linear regression	Yes; 1+ severe hypoglycaemic events led to a decrease in general health; following hospitalisation for hypoglycaemia, EQ‐5D Utility Index scores dropped (M Δ = –0.019, SE Δ = 0.024), p < 0.05
Haluzik et al. (2018)	23	NA (targeted question)	Yes; 1+ hypoglycaemic episodes impaired attendance at school or work; during the 1‐month study period, 2.5% of participants reported taking leave from school or work (M = 2.8 days), 2.7% reported arriving late and 4.9% reported leaving early as a direct consequence of hypoglycaemia
Jódar et al. (2020)	24	Linear regression	Mixed; 1+ hypoglycaemic episodes (of unspecified severity) led to larger improvements in physical health, but not mental health, following intervention; change in SF‐36 Physical Component scores was larger and more positive for those reporting hypoglycaemic episodes (M Δ = 1.04, SE Δ = 2.4) compared to those reporting no episodes (M Δ = 0.5, SE Δ = 0.2), p < 0.01, while change in SF‐36 Mental Component scores did not differ between those reporting episodes (M Δ = –0.5, SE Δ = 0.4) and those who did not (M Δ = 0.3, SE Δ = 0.2), p = 0.08
Malanda et al. (2011)	25	ANCOVA (adjusted for gender, age, education, diabetes duration, intervention)	Mixed; 1+ asymptomatic hypoglycaemic episodes led to increased perceived control over diabetes; pairwise comparisons showed changes in IPQ‐R Control subscale scores were larger and more positive for those reporting only asymptomatic episodes (M Δ = 1.04, SD Δ = 2.4), compared to those reporting self‐treated symptomatic episodes (M Δ = –0.3, SD Δ = 2.7), p = 0.007, d = 0.54, or no episodes (M Δ = –0.09, SD Δ = 3.3), p = 0.009, d = 0.37. There was no difference between those reporting symptomatic or no episodes, p > 0.05. Experiencing 1+ hypoglycaemic episodes did not affect general health, general well‐being or diabetes distress; three‐way comparisons for those reporting no episodes, only asymptomatic episodes or self‐treated symptomatic episodes showed no changes in EQ‐5D (M Δ = –0.04, 0.01, –0.03, SD Δ = 0.2, 0.2, 0.2), p = 0.23, W‐BQ12 (M Δ = –0.27, 0.03, 0.16, SD Δ = 5.2, 4.7, 4.0), p = 0.82 or IPQ‐R Emotion subscale scores (M Δ = 0.39, –0.27, –0.79, SD Δ = 3.6, 3.6, 3.9), p = 0.24
Nauck et al. (2019)	28	Linear regression (adjusted for gender, region, CVR, intervention)	Yes; 1+ severe hypoglycaemic events (requiring assistance or confirmed plasma glucose <3.1 mmol/L [56 mg/dl]) led to a decrease in general health; compared to those reporting no events, those reporting severe events experienced a drop in EQ‐5D Utility Index scores (M Δ = –0.018, SE Δ = 0.004), p < 0.001, but no change in VAS scores (M Δ = –0.009, SE Δ = 0.351), p = 0.98
Nicolucci et al. (2011)	29	Linear regression (adjusted for gender, age, HbA1c, weight, intervention)	Yes; 4+ hypoglycaemic episodes (of unspecified severity) led to a decrease in general well‐being and energy, and smaller improvements in mental health following intervention; compared to those reporting no episodes, those reporting 3+ episodes showed a drop in W‐BQ22 total, B = –5.41 (SE = 1.72), p = 0.002, and W‐BQ22 Energy subscale scores, B = –1.45 (SE = 0.49), p = 0.003, and smaller improvements in SF‐36 Mental Component scores, B = –5.03 (SE = 1.72), p = 0.004. Those reporting 1–3 (but not 4+) hypoglycaemic episodes showed a drop in W‐BQ22 Anxiety subscale scores, B = –1.76 (SE = 0.65), p = 0.007
Pathan et al. (2018)	38	NA (targeted question)	Yes; 1+ hypoglycaemic episodes impaired attendance at school or work; during the 1‐month study period, 3.2% of participants reported taking leave from school or work, 2.1% reported arriving late and 2.8% reported leaving early as a direct consequence of hypoglycaemia
Pichayapinyo et al. (2019)	31	Pearson's correlation	No; more frequent symptoms of hypoglycaemia did not lead to changes in depression, sleep disturbance, social support, diabetes distress or diabetes self‐efficacy; those reporting more symptoms showed a rise in PHQ‐8 total, r = 0.18, and PROMIS Sleep Disturbance subscale scores, r = 0.16, as well as a drop in SSQ total, r = –0.12, DDS total, r = –0.06, and SEDS total scores, r = –0.04, though none of these impacts were significant, ps > 0.05
Polonsky et al. (2018)	39	ANCOVA b (adjusted for gender, age, insulin status)	Mixed; 1+ self‐treated symptomatic hypoglycaemic episodes led to increased anxiety, diabetes distress and hypoglycaemic worry, but no change in depression or general well‐being; compared to those reporting no episodes, those reporting self‐treated episodes showed a rise in GAD total, β = 0.16, p < 0.01, DDS total, β = 0.12, p < 0.05, and HFS‐II worry subscale scores, β = 0.18, p < 0.01, but no change in PHQ‐8 total, β = 0.09, p > 0.05 or WHO‐5 total scores, β = 0.02, p > 0.05
Torre et al. (2019)	32	Linear regression b	No; 1+ self‐treated hypoglycaemic episodes or severe hypoglycaemic events did not lead to minimally important changes in general health; following self‐treated episodes or severe events, there was a non‐significant rise in EQ‐5D utility index, β = 0.29, p = 0.15, and EQ‐5D Visual Analogue Scale scores, β = 0.11, p = 0.57
Wieringa et al. (2018)	33	GEE (adjusted for gender, age, education, diabetes duration, HbA1c, BMI, and number of complications)	Mixed; 2+ self‐treated symptomatic hypoglycaemic episodes led to increased hypoglycaemic worry, but no change in general well‐being; compared to those reporting no episodes, those reporting 2+ self‐treated episodes showed a rise in HFS‐II Worry subscale scores, B = 1.33 (SE = 0.06), p < 0.001, but no change in WHO‐5 totals, B = −0.79 (SE = 0.95), p = 0.30. Experiencing 2+ severe hypoglycaemic events led to no change in hypoglycaemic worry or general well‐being; compared to those reporting no events, those reporting 2+ severe events showed no change in HFS‐II Worry subscale, B = 1.13 (SE = 0.12), p = 0.23, or WHO‐5 total scores, B = −1.63 (SE = 1.58), p = 0.31. A non‐significant interaction between time and hypoglycaemic events across all analyses suggested these impacts did not change over time
Yang et al. (2014)	34	Linear regression (adjusted for gender, age, BMI, diabetes duration, insulin history, HbA1c)	Mixed; 1+ severe hypoglycaemic events led to smaller improvements in general health following intervention; compared to those reporting no events, those reporting 1+ severe events showed smaller increases in EQ‐5D Visual Analogue Scale scores, B = 6.96, p < 0.001. When self‐treated symptomatic episodes and severe events were combined, this effect was no longer significant; following any hypoglycaemic episode, EQ‐5D Visual Analogue Scale scores did not change, B = 0.02, p = 0.96

Abbreviations: ANCOVA, Analysis of Covariance; CVR, Cardiovascular Risk; DDS, Diabetes Distress Scale; EQ‐5D, EuroQol 5‐Dimension health status instrument; GAD, General Anxiety Disorder scale; GEE, Generalised Estimating Equations; HFS‐II, Hypoglycaemia Fear Survey version II; IPQ‐R, Illness Perception Questionnaire Revised; PHQ‐8, Patient Health Questionnaire 8‐item; PROMIS, Patient‐Reported Outcomes Measurement Information System; SEDS, Self‐Efficacy for Diabetes Scale; SF‐36, Medical Outcomes Study Short Form 36‐item health survey; SSQ, Social Support Questionnaire; W‐BQ12 and W‐BQ22; Well‐Being Questionnaire 12‐item and 22‐item; WHO‐5, World Health Organisation 5‐item well‐being index.

^a Studies which did not report estimates of effect size were omitted from this table.

^b To facilitate comparisons across studies, OR and 95% CI values were converted to β and p, respectively.