Table 1:

Summary of evidence of harms associated with screening for type 2 diabetes^*

No. of studies	utcome measure	Mean score ± SD		Absolute effect (95% CI)^†	GRADE quality of evidence

		No invitation to screening	Invitation to screening
	Anxiety
1 RCT18 n = 355	Spielberger State Anxiety Inventory	6 wk after last contact^‡		Mean score 3.5 higher (0.22 to 6.78)	Moderate^§^¶^**
1 RCT18 n = 355	Spielberger State Anxiety Inventory	34.1 ± 12.1 n = 168	37.6 ± 12.2 n = 77	Mean score 3.5 higher (0.22 to 6.78)	Moderate^§^¶^**

	Anxiety	At baseline^††		Mean score 0.53 lower (−2.60 to 1.54)	Low^¶^**^‡‡^§§
1 RCT19 n = 7380	Spielberger State Anxiety Inventory	32.7 ± 11.5 n = 199	32.7 ± 11.6 n = 2 468	Mean score 0.53 lower (−2.60 to 1.54)

		At 3–6 mo		Mean score 1.51 higher (−0.17 to 3.20)
		31.8 ± 11.4 n = 358	33.5 ± 12.0 n = 2 504	Mean score 1.51 higher (−0.17 to 3.20)

		At 12–15 mo		Mean score 0.57 higher (−1.11 to 2.24)
		32.8 ± 11.8 n = 304	35.5 ± 12.2 n = 2 377	Mean score 0.57 higher (−1.11 to 2.24)

	Hospital Anxiety and Depression Scale: Anxiety Subscale	At baseline^††		Mean score 0.46 lower (−0.99 to 0.07)	Low^§^¶^**^‡‡^§§
		6.42 ± 4.39 n = 255	6.04 ± 3.79 n = 3 140	Mean score 0.46 lower (−0.99 to 0.07)

		At 3–6 mo		Mean score 0.12 lower (−0.55 to 0.32)
		5.97 ± 3.86 n = 442	5.91 ± 3.89 n = 3 159	Mean score 0.12 lower (−0.55 to 0.32)

		At 12–15 mo		Mean score 0.01 lower (−0.47 to 0.45)
		5.81 ± 3.87 n = 377	5.85 ± 3.87 n = 3 034	Mean score 0.01 lower (−0.47 to 0.45)

	Depression
	Hospital Anxiety and Depression Scale: Depression Subscale	At baseline^††		Mean score 0.37 lower (−0.93 to 0.18)	Low^¶^**^‡‡^§§
		4.52 ± 3.48 n = 256	4.24 ± 3.31 n = 3 161	Mean score 0.37 lower (−0.93 to 0.18)

		At 3–6 mo		Mean score 0.01 higher (−0.51 to 0.54)
		4.18 ± 3.38 n = 444	4.24 ± 3.40 n = 3 177	Mean score 0.01 higher (−0.51 to 0.54)

		At 12–15 mo		Mean score 0.22 higher (−0.31 to 0.74)
		4.03 ± 3.35 n = 378	4.28 ± 3.40 n = 3 049	Mean score 0.22 higher (−0.31 to 0.74)

Note: CI = confidence interval, RCT = randomized controlled trial, SD = standard deviation.
↵* ur systematic review of harms associated with screening for type 2 diabetes in adults of any age identified 2 RCTs.
↵† Eborall et al.19 used adjusted mean differences for age and comorbidity (use of antihypertensives) to compute absolute effect.
↵‡ Questionnaire was sent 6 weeks after last contact (either test or invitation).
↵§ Unclear allocation concealment.
↵¶ No information regarding blinding.
↵** Quality rating is for a single study; thus, imprecision and publication bias criteria were rated as “no” and “unlikely.”
↵†† Questionnaire was given immediately after the initial blood test for those who attended screening, or after first contact for controls; data for those who attended screening were included in the analysis only if the questionnaire was completed and returned before the results of the test were received.
↵‡‡ A nonrandomized sample of screening practices was used.
↵§§ Large loss to follow-up (for the follow-up periods 3–6 and 12–15 mo).