Local Mortality: India states

[st2048]

Just wanted to share some data I've found from Indian states/UTs that I haven't seen in the most common repositories, in case this is of interest for local or preliminary mortality. Some official data for the whole of 2021 is starting to appear. Happy to share other data as I find it if you find it useful.

Goa

• Full monthly data from 2015 to 2021. Matches exactly with national CRS reports.
• Source: State Annual Report for Births & Deaths
• When compared to the SRS data for 2015-20, implied completeness is 132, 129, 138, 144, 152 and 159 per cent respectively. As Goa is a relatively small state, some of this difference might be because SRS is by place of residence whereas CRS is by place of death - however, as I am sure you know there are big states as well where the SRS underestimates mortality, so that could also be the reason here.
• Goa has a reputation of having a strong civil registration system and is the only state with 100% registration according to NFHS-5 (the survey in Goa was conducted in 2019 and asked about percentage of deaths from the past 3 years registered with the civil authority).
• Using 2015-19 trend as baseline would imply around 320-330 excess deaths per 100k.

State	Year	Month	Deaths
Goa	2015	1	1228
Goa	2015	2	1065
Goa	2015	3	1101
Goa	2015	4	961
Goa	2015	5	989
Goa	2015	6	1020
Goa	2015	7	1083
Goa	2015	8	1062
Goa	2015	9	1078
Goa	2015	10	1112
Goa	2015	11	1025
Goa	2015	12	1014
Goa	2016	1	863
Goa	2016	2	1041
Goa	2016	3	1017
Goa	2016	4	996
Goa	2016	5	1064
Goa	2016	6	1110
Goa	2016	7	1089
Goa	2016	8	1286
Goa	2016	9	1126
Goa	2016	10	1084
Goa	2016	11	1181
Goa	2016	12	1273
Goa	2017	1	1225
Goa	2017	2	1072
Goa	2017	3	1128
Goa	2017	4	986
Goa	2017	5	1028
Goa	2017	6	948
Goa	2017	7	1107
Goa	2017	8	1099
Goa	2017	9	1245
Goa	2017	10	1049
Goa	2017	11	1122
Goa	2017	12	1011
Goa	2018	1	1388
Goa	2018	2	1067
Goa	2018	3	1112
Goa	2018	4	1111
Goa	2018	5	1044
Goa	2018	6	935
Goa	2018	7	1100
Goa	2018	8	1072
Goa	2018	9	1072
Goa	2018	10	1169
Goa	2018	11	1076
Goa	2018	12	926
Goa	2019	1	1373
Goa	2019	2	1129
Goa	2019	3	1013
Goa	2019	4	1098
Goa	2019	5	1063
Goa	2019	6	896
Goa	2019	7	1149
Goa	2019	8	1410
Goa	2019	9	1170
Goa	2019	10	1271
Goa	2019	11	1084
Goa	2019	12	1195
Goa	2020	1	1259
Goa	2020	2	1099
Goa	2020	3	905
Goa	2020	4	1056
Goa	2020	5	1155
Goa	2020	6	1006
Goa	2020	7	1097
Goa	2020	8	1217
Goa	2020	9	1670
Goa	2020	10	1538
Goa	2020	11	1297
Goa	2020	12	1302
Goa	2021	1	1166
Goa	2021	2	1121
Goa	2021	3	1115
Goa	2021	4	1428
Goa	2021	5	3440
Goa	2021	6	2320
Goa	2021	7	1404
Goa	2021	8	1317
Goa	2021	9	1313
Goa	2021	10	1354
Goa	2021	11	1201
Goa	2021	12	1158

Delhi

• Full monthly data from 2015 to 2021. Matches exactly with national CRS reports.
• Source: Annual Report on Registration of Births and Deaths in Delhi
• When compared to the SRS data for 2015-20, implied completeness is 188, 188, 192, 225, 228 and 195 per cent respectively. In the case of Delhi, my understanding is that this difference is largely explained by non-residents who use the city hospitals.
• According to NFHS-5 registration completeness for Delhi residents was 89% (the survey in Delhi was mostly conducted in 2020 and asked about percentage of deaths from the past 3 years registered with the civil authority).
• When looking at the plotted monthly data, there does seem to be a drop from March 2020 onwards - whilst a 2015-19 trend baseline works well for the first couple of months in 2020, it does seem a bit too high after that, as it implies significantly below baseline mortality throughout 2020 and 2021 except for the months affected by covid waves.
• Back when The Hindu had published their story about Delhi they had reported a drop in patients coming from other states in 2020, which could be the reason behind this drop - based on the plotted monthly data it does seem this trend continued throughout 2021. At the same time, I guess it's also possible that some people who died in Delhi hospitals during the covid peaks may have been residents from other states.
• The reports mention the data is by month of occurrence. The monthly patterns do match with the covid waves, and it is interesting to see how the 2021 peak in Delhi was in April as opposed to May, suggesting the wave hit Delhi earlier than some other parts of the country.

State	Year	Month	Deaths
Delhi	2015	1	12412
Delhi	2015	2	11445
Delhi	2015	3	10767
Delhi	2015	4	9400
Delhi	2015	5	10080
Delhi	2015	6	8884
Delhi	2015	7	9094
Delhi	2015	8	10385
Delhi	2015	9	11254
Delhi	2015	10	10362
Delhi	2015	11	9887
Delhi	2015	12	10546
Delhi	2016	1	13607
Delhi	2016	2	11336
Delhi	2016	3	10834
Delhi	2016	4	10314
Delhi	2016	5	11093
Delhi	2016	6	9905
Delhi	2016	7	11724
Delhi	2016	8	12458
Delhi	2016	9	14227
Delhi	2016	10	12431
Delhi	2016	11	11300
Delhi	2016	12	12403
Delhi	2017	1	12811
Delhi	2017	2	10568
Delhi	2017	3	11052
Delhi	2017	4	10569
Delhi	2017	5	11182
Delhi	2017	6	10309
Delhi	2017	7	10656
Delhi	2017	8	11995
Delhi	2017	9	10956
Delhi	2017	10	12130
Delhi	2017	11	11687
Delhi	2017	12	12202
Delhi	2018	1	15745
Delhi	2018	2	12565
Delhi	2018	3	12053
Delhi	2018	4	11073
Delhi	2018	5	11203
Delhi	2018	6	11424
Delhi	2018	7	10456
Delhi	2018	8	11562
Delhi	2018	9	12481
Delhi	2018	10	12316
Delhi	2018	11	11630
Delhi	2018	12	13025
Delhi	2019	1	15565
Delhi	2019	2	12990
Delhi	2019	3	12423
Delhi	2019	4	10802
Delhi	2019	5	11153
Delhi	2019	6	11531
Delhi	2019	7	10956
Delhi	2019	8	11754
Delhi	2019	9	11550
Delhi	2019	10	11639
Delhi	2019	11	11492
Delhi	2019	12	13429
Delhi	2020	1	14812
Delhi	2020	2	13203
Delhi	2020	3	11064
Delhi	2020	4	8565
Delhi	2020	5	11631
Delhi	2020	6	15364
Delhi	2020	7	10014
Delhi	2020	8	9854
Delhi	2020	9	10981
Delhi	2020	10	10046
Delhi	2020	11	13415
Delhi	2020	12	13840
Delhi	2021	1	12273
Delhi	2021	2	10044
Delhi	2021	3	10691
Delhi	2021	4	35291
Delhi	2021	5	27346
Delhi	2021	6	9958
Delhi	2021	7	9663
Delhi	2021	8	11629
Delhi	2021	9	10707
Delhi	2021	10	11349
Delhi	2021	11	10791
Delhi	2021	12	11734

Andhra Pradesh

I'm not sure whether this is the same source as the one you have for your Andhra Pradesh data - I recently scrapped it again so I'm leaving here this more updated version in case it's useful.

• Monthly data from January 2018 to September 2022 - I excluded October and November 2022 because there may be significant revisions.
• Source: Online Civil Registration Portal - I initially found this link here. I am using 'registered deaths' column.
• This is online registration only, and represents 88% of all registered deaths in 2018, 90% in 2019 and 94% in 2020, when compared to the final numbers for the state in the CRS national reports.
• When compared to the SRS data for 2015-20, implied completeness (of the full CRS data, not just the online version) is 86, 90, 95, 108, 120 and 137 per cent respectively. Andhra Pradesh is a large state so I doubt that deaths of non-residents are a significant factor in these numbers, plus several other large states in South India have 100+ completeness according to the SRS, suggesting the SRS is underestimating mortality.
• The SRS suggests that the increase in deaths from 2019 to 2020 in Andhra Pradesh was completely explained by an increase in registration (it estimates 334,816 deaths for 2019 and 331,374 for 2020). The monthly data suggests this is implausible, as deaths only rise in the months were covid was prevalent and then fall back again. The data is by month of registration but I also recently analysed the column on 'registered within SLA (3 days)' and the monthly pattern is broadly similar, so the increase in late 2020 cannot be just late registrations.
• According to NFHS-5 registration completeness for Andhra Pradesh was 80.2% (the survey in Andhra Pradesh was conducted in 2019 and asked about percentage of deaths from the past 3 years registered with the civil authority). If I assume that two thirds of the deaths in NFHS-5 were from 2017 and 2018, with the remaining third from parts of 2016 and 2019, the equivalent SRS completeness estimate for that period would be 103%. This would suggest that if NFHS-5 figures are correct the SRS may be underestimating mortality by more than 20 percentage points in the state, though this is just a rough calculation.
• If we rescale SRS completeness data according to this estimate, we would get 67, 70, 75, 84, 94 and 107 per cent for 2015-20. This would suggest that the state may be getting close to completeness from 2019 onwards. I would be cautious with using SRS year on year changes though, as in the same way that the SRS seems to be wrong when it estimates that the rise from 2019 to 2020 is exclusively due to changes in completeness, it is possible that the SRS may also be wrong when it suggests that the rises in previous years were also mostly due to changes in completeness. I guess it is hard to know.
• Regardless of the specific baseline, it seems clear to me that there has been a big number of excess deaths in the state, and the peaks match the covid waves.

State	Year	Month	Deaths
Andhra Pradesh	2018	1	33457
Andhra Pradesh	2018	2	28635
Andhra Pradesh	2018	3	27174
Andhra Pradesh	2018	4	25831
Andhra Pradesh	2018	5	24226
Andhra Pradesh	2018	6	26848
Andhra Pradesh	2018	7	26293
Andhra Pradesh	2018	8	27761
Andhra Pradesh	2018	9	26483
Andhra Pradesh	2018	10	28348
Andhra Pradesh	2018	11	26645
Andhra Pradesh	2018	12	29798
Andhra Pradesh	2019	1	32149
Andhra Pradesh	2019	2	31050
Andhra Pradesh	2019	3	27865
Andhra Pradesh	2019	4	26212
Andhra Pradesh	2019	5	29813
Andhra Pradesh	2019	6	30099
Andhra Pradesh	2019	7	28360
Andhra Pradesh	2019	8	28563
Andhra Pradesh	2019	9	30341
Andhra Pradesh	2019	10	33559
Andhra Pradesh	2019	11	33095
Andhra Pradesh	2019	12	31115
Andhra Pradesh	2020	1	31890
Andhra Pradesh	2020	2	26731
Andhra Pradesh	2020	3	25332
Andhra Pradesh	2020	4	22340
Andhra Pradesh	2020	5	27987
Andhra Pradesh	2020	6	28275
Andhra Pradesh	2020	7	34583
Andhra Pradesh	2020	8	54757
Andhra Pradesh	2020	9	55556
Andhra Pradesh	2020	10	41913
Andhra Pradesh	2020	11	38472
Andhra Pradesh	2020	12	39375
Andhra Pradesh	2021	1	38170
Andhra Pradesh	2021	2	32160
Andhra Pradesh	2021	3	33868
Andhra Pradesh	2021	4	39908
Andhra Pradesh	2021	5	136548
Andhra Pradesh	2021	6	67576
Andhra Pradesh	2021	7	39491
Andhra Pradesh	2021	8	35804
Andhra Pradesh	2021	9	38772
Andhra Pradesh	2021	10	35496
Andhra Pradesh	2021	11	35414
Andhra Pradesh	2021	12	37707
Andhra Pradesh	2022	1	37729
Andhra Pradesh	2022	2	37433
Andhra Pradesh	2022	3	32102
Andhra Pradesh	2022	4	28649
Andhra Pradesh	2022	5	29898
Andhra Pradesh	2022	6	32060
Andhra Pradesh	2022	7	32122
Andhra Pradesh	2022	8	35476
Andhra Pradesh	2022	9	35204

[akarlinsky]

This is very helpful. The Delhi 2021 one is really fresh (I have been accessing that page every once in a while and last time I looked there was only 2020) and Goa I looked for but couldn't find.

Please send me an email at your earliest convenience. Address is on my personal website: https://akarlinsky.github.io/

[st2048]

Please send me an email at your earliest convenience. Address is on my personal website: https://akarlinsky.github.io/

Thanks - just emailed you!

[dkobak]

@st2048 @akarlinsky Can this issue be closed now?