Table 2.

Poisson segmented regression models of the impact of COVID-19 lockdown on HIV services at 65 clinics in KwaZulu-Natal, South Africa

		Incidence rate ratio at lockdown	Incidence rate ratio at study end	Pre-lockdown trend*	Post-lockdown trend*
HIV testing†
Overall		0·524 (0·446–0·615)	0·827 (0·704–0·972)	1·015 (1·009–1·022)	1·183 (1·113–1·256)
Gender
	Women	0·563 (0·480–0·661)	0·865 (0·741–1·010)	1·013 (1·007–1·019)	1·169 (1·100–1·247)
	Men	0·456 (0·387–0·537)	0·761 (0·638–0·907)	1·020 (1·013–1·028)	1·210 (1·147–1·277)
District
	Urban	0·515 (0·436–0·610)	0·827 (0·700–0·978)	1·016 (1·010–1·023)	1·190 (1·115–1·270)
	Rural	0·823 (0·681–0·994)	0·785 (0·639–0·965)	0·996 (0·985–1·006)	0·980 (0·922–1·042)
ART initiation‡
Overall		0·538 (0·459–0·630)	0·753 (0·637–0·890)	0·996 (0·991–1·004)	1·156 (1·085–1·230)
Gender
	Women	0·495 (0·417–0·588)	0·801 (0·651–0·985)	0·996 (0·987–1·004)	1·225 (1·118–1·341)
	Men	0·616 (0·513–0·740)	0·675 (0·571–0·797)	1·000 (0·996–1·009)	1·040 (0·957–1·132)
Age, years
	0–14	0·370 (0·153–0·896)	0·544 (0·270–1·096)	0·996 (0·987–1·004)	1·175 (0·640–2·164)
	15–24	0·400 (0·334–0·479)	0·772 (0·640–0·932)	1·022 (1·013–1·031)	1·363 (1·256–1·476)
	25–49	0·577 (0·492–0·676)	0·739 (0·624–0·876)	0·991 (0·987–1·000)	1·108 (1·040–1·180)
	≥50	0·481 (0·399–0·579)	0·796 (0·664–0·953)	0·991 (0·983–1·000)	1·230 (1·161–1·335)
District
	Urban	0·536 (0·458–0·628)	0·750 (0·631–0·892)	0·996 (0·991–1·004)	1·156 (1·080–1·230)
	Rural	0·588 (0·365–0·947)	0·863 (0·625–1·191)	0·987 (0·983–0·996)	1·166 (0·869–1·560)
ART collection visits‡
Overall		0·932 (0·794–1·093)	0·859 (0·747–0·989)	1·009 (1·000–1·013)	0·974 (0·900–1·053)
Gender
	Women	0·916 (0·780–1·075)	0·865 (0·753–0·994)	1·009 (1·000–1·013)	0·983 (0·908–1·062)
	Men	0·965 (0·823–1·131)	0·847 (0·733–0·978)	1·009 (1·004–1·013)	0·953 (0·880–1·031)
Age, years
	0–14	0·908 (0·790–1·043)	0·780 (0·703–0·866)	1·013 (1·009–1·017)	0·949 (0·888–1·017)
	15–24	0·875 (0·766–1·001)	0·851 (0·753–0·960)	1·026 (1·017–1·031)	1·013 (0·957–1·071)
	25–49	0·933 (0·796–1·094)	0·865 (0·751–0·995)	1·009 (1·000–1·013)	0·974 (0·900–1·053)
	≥50	0·946 (0·787–1·137)	0·836 (0·710–0·984)	1·004 (1·000–1·009)	0·949 (0·861–1·049)
District
	Urban	0·926 (0·786–1·090)	0·850 (0·736–0·982)	1·009 (1·004–1·013)	0·970 (0·896–1·053)
	Rural	1·065 (0·969–1·171)	1·080 (0·985–1·185)	0·996 (0·991–1·000)	1·000 (0·957–1·044)
Missed ART collection visits‡
Overall		1·926 (1·585–2·341)	0·682 (0·455–1·021)	1·000 (0·996–1·004)	0·569 (0·432–0·747)
Gender
	Women	2·045 (1·686–2·481)	0·691 (0·454–1·052)	0·996 (0·991–1·000)	0·555 (0·419–0·734)
	Men	1·710 (1·399–2·090)	0·664 (0·457–0·965)	1·000 (0·996–1·004)	0·601 (0·463–0·779)
Age, years
	0–14	1·568 (1·325–1·856)	0·555 (0·429–0·717)	1·004 (1·000–1·009)	0·572 (0·472–0·690)
	15–24	1·682 (1·413–2·001)	0·554 (0·374–0·820)	1·022 (1·017–1·026)	0·561 (0·430–0·727)
	25–49	1·947 (1·594–2·378)	0·685 (0·451–1·042)	1·000 (0·996–1·000)	0·566 (0·428–0·751)
	≥50	2·030 (1·665–2·474)	0·782 (0·552–1·109)	0·987 (0·983–0·991)	0·589 (0·463–0·751)
District
	Urban	1·991 (1·621–2·446)	0·677 (0·433–1·057)	1·000 (0·996–1·004)	0·555 (0·412–0·751)
	Rural	1·274 (1·111–1·460)	0·745 (0·636–0·872)	1·004 (0·996–1·009)	0·751 (0·674–0·834)

Data are incidence rate ratio (95% CI) or trend (95% CI). ART=antiretroviral therapy.

^*Slope change per month

^†Autocorrelation addressed using Newey–West standard errors to calculate CI, with lag up to 2.

^‡Autocorrelation addressed using Newey–West standard errors to calculate CI, with lag up to 3.