Global surveillance of trends in cancer survival 2000–14 (CONCORD-3): analysis of individual records for 37 513 025 patients diagnosed with one of 18 cancers from 322 population-based registries in 71 countries

May combine two claims.

CONCORD-3 covers patients diagnosed in 2000-2014 with one of 18 cancers that account for 75% of the global cancer burden.

“CONCORD-3 includes data for patients diagnosed during 2000–2014 with one of 18 malignancies that represent 75% of the global cancer burden (Table 1).”

Classified as probably not a substantial claim.

Eighteen cancers or cancer groups were covered.

“The study includes 18 cancers or groups of cancers: oesophagus, stomach, colon, rectum, liver, pancreas, lung, breast (women), cervix, ovary, prostate and melanoma of the skin in adults, and brain tumours, leukaemias and lymphomas in both adults and children.”

May combine two claims.

Survival generally improved over time, including for some cancers with high lethality.

“Survival trends are generally increasing, even for some of the more lethal cancers:”

Probably not a substantial claim: may not pass S3 (load-bearing).

Eight countries joined global cancer survival surveillance for the first time: Morocco; Costa Rica, Mexico (children) and Peru; Iran, Kuwait and Singapore; and Greece (children).

“Eight countries are participating in the global surveillance of cancer survival for the first time: Africa: Morocco; Central and South America: Costa Rica (national), Mexico (children, national), and Peru; Asia: Iran, Kuwait (national) and Singapore (national), and Europe: Greece (children, national).”

May combine two claims.

New ICD-O-3 entities for acute lymphoblastic leukaemia/lymphoma were unused by registries in 42 of 58 countries reporting childhood ALL for 2010-2014, rare in eight countries, but ranged from 11% to 89% in large datasets from Australia, Belgium, Canada, the Netherlands, Puerto Rico, Singapore, Taiwan and the US.

“These new entities were not used at all by registries in 42 of the 58 countries that submitted data for children diagnosed with ALL during 2010–2014, and very rarely in eight countries (combined number of children coded to a new entity below 100), but the proportions ranged from 11% to 89% in large data sets from Australia, Belgium, Canada, the Netherlands, Puerto Rico, Singapore, Taiwan and the U”

May combine two claims.

The study received data from 322 population-based registries across 71 countries and territories; 47 of these had complete population coverage.

“Data were provided by 322 population-based cancer registries in 71 countries and territories, of which 47 provided data with 100% population coverage.”

May combine two claims.

Analysis used data from 322 registries in 71 countries and territories, representing 81% of 400 operational invited registries, for patients diagnosed in 2000-2014 with vital status at least five years after diagnosis or at 31 December 2014.

“We analysed data provided by 322 cancer registries (81% of the 400 operational registries invited) in 71 countries and territories (text-figure 1), for patients diagnosed with cancer during the 15-year period 2000–2014, with data on their vital status at least five years after diagnosis, or at 31 December 2014.”

May combine two claims.

Breast cancer survival varied widely between countries, reaching as low as 40% in South Africa.

“but international differences remain very wide, with levels as low as 40% in South Africa.”

May combine two claims.

In 47 countries, registry data covered the entire national population; 41 covered both adults and children, and 6 covered children only.

“For 47 countries, data were provided with 100% coverage of the national population: 41 for both adults and children, and 6 for children only (Argentina, Belarus, France, Greece, Mexico and Switzerland) (Table 3).”

May combine two claims.

Oesophageal cancer survival was highest in Japan (36%), China (34%) and Korea* (31%), and below 30% in all other countries.

“Survival was highest in Japan (36%), China (34%) and Korea* (31%), and below 30% in all other countries (Table 5, web-figure 2).”

• AssumptionRests on an assumption: Patients lost to follow-up or censored alive have the same subsequent survival as patients with complete follow-up.
• AssumptionRests on an assumption: For the 34 registries without reliable local all-cause mortality data, UN national life tables are adequate proxies for the registry territory's background mortality.
• AssumptionRests on an assumption: Exclusion of patients registered only from a death certificate or at autopsy does not substantially bias the survival estimates.

May combine two claims.

Between 1995-1999 and 2000-2014, rectal cancer survival trends were generally increasing, especially since 2000, in Asia, Europe and Oceania; survival increased by 5-10% in Japan, Taiwan*, Finland*, Italy, Sweden and New Zealand*; by 10% or more in Canada, China, Israel*, Korea*, 13 European countries and Australia*; and by about 20% in Korea* and Slovenia*.

“Survival trends between 1995–1999 and 2000–2014 were generally increasing, especially since 2000, in Asia, Europe and Oceania (web-figure 3). Survival increased by 5–10% in Japan and Taiwan*; in Finland*, Italy and Sweden; and in New Zealand*. Over the same period, 5-year survival increased by 10% or more in Canada; in China, Israel* and Korea*; in 13 European countries (North: Denmark*, Estonia*,”

• AssumptionRests on an assumption: Patients lost to follow-up or censored alive have the same subsequent survival as patients with complete follow-up.
• AssumptionRests on an assumption: For the 34 registries without reliable local all-cause mortality data, UN national life tables are adequate proxies for the registry territory's background mortality.
• AssumptionRests on an assumption: Exclusion of patients registered only from a death certificate or at autopsy does not substantially bias the survival estimates.

May combine two claims.

Rectal cancer survival was below 50% in Slovakia* (49%), Poland* (48%), Croatia* (48%), Bulgaria* (46%), Ecuador (45%), Thailand (44%), Russia (42%) and India (30%).

“Five-year survival was below 50% in Slovakia* (49%), Poland* (48%), Croatia* (48%), Bulgaria* (46%), Ecuador (45%), Thailand (44%), Russia (42%) and India (30%).”

• AssumptionRests on an assumption: Patients lost to follow-up or censored alive have the same subsequent survival as patients with complete follow-up.
• AssumptionRests on an assumption: For the 34 registries without reliable local all-cause mortality data, UN national life tables are adequate proxies for the registry territory's background mortality.
• AssumptionRests on an assumption: Exclusion of patients registered only from a death certificate or at autopsy does not substantially bias the survival estimates.

May combine two claims.

Between 1995-1999 and 2000-2014, colon cancer survival trends were generally flat or increasing; survival increased by 5-10% in 11 countries, and by more than 10% in China, Israel*, Korea* and eight European countries.

“Survival trends between 1995–1999 and 2000–2014 were generally flat or increasing (web-figure 3). Five-year survival increased over this period by 5–10% in 11 countries: in Canada; in Japan and Taiwan*, and in 7 European countries (North: Estonia*, Ireland*, Lithuania*, Norway*, Sweden*; South: Italy; West: the Netherlands*) and in Australia*. Over the same period, survival increased by more than”

• AssumptionRests on an assumption: Patients lost to follow-up or censored alive have the same subsequent survival as patients with complete follow-up.
• AssumptionRests on an assumption: For the 34 registries without reliable local all-cause mortality data, UN national life tables are adequate proxies for the registry territory's background mortality.
• AssumptionRests on an assumption: Exclusion of patients registered only from a death certificate or at autopsy does not substantially bias the survival estimates.

May combine two claims.

Age-standardised five-year net survival for stomach cancer was generally 20-40%, with very wide variation in Asia.

“Age-standardised five-year net survival was generally in the range 20–40%, with very wide variation in Asia (web-figure 4).”

• AssumptionRests on an assumption: Patients lost to follow-up or censored alive have the same subsequent survival as patients with complete follow-up.
• AssumptionRests on an assumption: For the 34 registries without reliable local all-cause mortality data, UN national life tables are adequate proxies for the registry territory's background mortality.
• AssumptionRests on an assumption: Exclusion of patients registered only from a death certificate or at autopsy does not substantially bias the survival estimates.

Probably not a substantial claim: may not pass S3 (load-bearing).

The study received more than 4,700 data sets.

“We received more than 4,700 data sets.”

Probably not a substantial claim: may not pass S3 (load-bearing).

For 207 registries, life table construction depended on whether raw data or mortality rates were received and whether they were by single year of age or five-year age group.

“For 207 registries, the method of life table construction depended on whether we received raw data (numbers of deaths and populations) or mortality rates, and on whether the raw data or the mortality rates were by single year of age (“complete”) or by five-year age group (“abridged”).”

May combine two claims.

Colon cancer survival was below 50% in Ecuador (48%), Thailand (47%), Russia (45%) and India (39%).

“Five-year net survival was below 50% in Ecuador (48%), Thailand (47%), Russia (45%) and India (39%).”

• AssumptionRests on an assumption: Patients lost to follow-up or censored alive have the same subsequent survival as patients with complete follow-up.
• AssumptionRests on an assumption: For the 34 registries without reliable local all-cause mortality data, UN national life tables are adequate proxies for the registry territory's background mortality.
• AssumptionRests on an assumption: Exclusion of patients registered only from a death certificate or at autopsy does not substantially bias the survival estimates.

Probably not a substantial claim: may not pass S3 (load-bearing).

More than 80 cancer registries joined the CONCORD programme for the first time.

“More than 80 cancer registries joined the CONCORD programme for the first time.”

Classified as probably not a substantial claim. May combine two claims.

For 109 registries, death and population counts were obtained from the registry or national statistical authority, and life tables were derived for 2001 and 2013 if possible, each centred on three calendar years.

“For 109 registries, we obtained death and population counts from the registry, or the relevant national statistical authority. We derived life tables for 2001 and 2013 if possible, each centred on three calendar years of data (e.g. 2000–2002, 2012–2014) to increase the robustness of the rates.”

May combine two claims.

Rectal cancer survival was higher than 70% in Jordan* (73%), Korea* (71%) and Australia* (71%).

“Survival was higher than 70% in Jordan* (73%), Korea* (71%), and Australia* (71%).”

• AssumptionRests on an assumption: Patients lost to follow-up or censored alive have the same subsequent survival as patients with complete follow-up.
• AssumptionRests on an assumption: For the 34 registries without reliable local all-cause mortality data, UN national life tables are adequate proxies for the registry territory's background mortality.
• AssumptionRests on an assumption: Exclusion of patients registered only from a death certificate or at autopsy does not substantially bias the survival estimates.

May combine two claims.

Stomach cancer survival was 20-29% in 25 countries and below 20% in Chile, Ecuador, Thailand and Bulgaria*.

“Survival was in the range 20–29% in 25 countries (Mauritius*, Brazil, Kuwait*, Turkey, 20 European countries, New Zealand*), and below 20% in Chile, Ecuador, Thailand and Bulgaria*.”

• AssumptionRests on an assumption: Patients lost to follow-up or censored alive have the same subsequent survival as patients with complete follow-up.
• AssumptionRests on an assumption: For the 34 registries without reliable local all-cause mortality data, UN national life tables are adequate proxies for the registry territory's background mortality.
• AssumptionRests on an assumption: Exclusion of patients registered only from a death certificate or at autopsy does not substantially bias the survival estimates.

Probably not a substantial claim: may not pass S3 (load-bearing).

For 120 registries, complete life tables were obtained without interpolation or smoothing for each calendar year 2000-2014.

“For 120 registries, we obtained complete life tables that did not require interpolation or smoothing, for each calendar year 2000–2014.”

May combine two claims.

Between 1995-1999 and 2010-2014, stomach cancer survival trends were rather flat in most countries, but after 2000 increased by 5% in Canada, the US, Singapore* and six European countries; by up to 10% in Israel*, Japan, Estonia* and Ireland*; and by 20% or more in China and Korea*.

“Survival trends between 1995–1999 and 2010–2014 were rather flat in most countries, but after 2000, survival increased by 5% in Canada and the US, Singapore* and in 6 European countries (North: Denmark*, Lithuania*, the UK*; East: Poland*; West: Austria*, the Netherlands*). Over the same 20 years, 5-year survival increased by up to 10% in Israel*, Japan, Estonia* and Ireland*, and by 20% or more”

• AssumptionRests on an assumption: Patients lost to follow-up or censored alive have the same subsequent survival as patients with complete follow-up.
• AssumptionRests on an assumption: For the 34 registries without reliable local all-cause mortality data, UN national life tables are adequate proxies for the registry territory's background mortality.
• AssumptionRests on an assumption: Exclusion of patients registered only from a death certificate or at autopsy does not substantially bias the survival estimates.

Probably not a substantial claim: may not pass S3 (load-bearing).

Of the examined registrations, 2,701,998 (6.4%) were for in situ cancer.

“Of these, 2,701,998 (6.4%) were for an in situ cancer, mostly of the cervix (54.6% of 1,708,084 women), breast (10.6% of 7,377,462 women), colon (4.4% of 4,619,536 adults) or prostate (0.6% of 6,069,630 men) (web-table 2).”

May combine two claims.

Rectal cancer survival varied widely, and the range of survival estimates in Asia for 2010-2014 was even wider than for colon cancer.

“As for colon, five-year net survival for rectal cancer varied widely. The range of survival estimates in Asia for 2010–2014 was even wider than for colon cancer (web-figure 4).”

• AssumptionRests on an assumption: Patients lost to follow-up or censored alive have the same subsequent survival as patients with complete follow-up.
• AssumptionRests on an assumption: For the 34 registries without reliable local all-cause mortality data, UN national life tables are adequate proxies for the registry territory's background mortality.
• AssumptionRests on an assumption: Exclusion of patients registered only from a death certificate or at autopsy does not substantially bias the survival estimates.

Probably not a substantial claim: may not pass S3 (load-bearing).

For the 42 participating US states, life tables by state, race and socioeconomic status were provided by the US National Cancer Institute.

“For the 42 participating states in the US, we used life tables by state, race and socioeconomic status, provided by the US National Cancer Institute (Mariotto A, NCI, personal communication, 26 January 2016).”

May combine two claims.

Between 2000 and 2014, 5-year net survival for oesophageal cancer increased by 4-5% in Singapore*, three European countries and the US; by 6-10% in China, Israel*, Japan, six European countries and Australia*; and by 13% in Korea*.

“Survival trends from 2000 to 2014 increased by 4–5% in Singapore*, 3 European countries (North: Denmark*, the UK*; West: Germany) and in the US (web-figure 3). Survival increased by 6–10% in China, Israel* and Japan, in 6 European countries (North: Ireland*, Norway*; South: Portugal*; West: Belgium*, the Netherlands*, Switzerland) and in Australia*. The increase in Korea* was 13%.”

• AssumptionRests on an assumption: Patients lost to follow-up or censored alive have the same subsequent survival as patients with complete follow-up.
• AssumptionRests on an assumption: For the 34 registries without reliable local all-cause mortality data, UN national life tables are adequate proxies for the registry territory's background mortality.
• AssumptionRests on an assumption: Exclusion of patients registered only from a death certificate or at autopsy does not substantially bias the survival estimates.

May combine two claims.

In 2010-2014, 5-year age-standardised net survival for oesophageal cancer was 10-30% in most countries, with a wider range in Asia.

“In 2010–2014, 5-year age-standardised net survival was in the range 10–30% in most countries, with a much wider range in Asia (web-figure 4).”

• AssumptionRests on an assumption: Patients lost to follow-up or censored alive have the same subsequent survival as patients with complete follow-up.
• AssumptionRests on an assumption: For the 34 registries without reliable local all-cause mortality data, UN national life tables are adequate proxies for the registry territory's background mortality.
• AssumptionRests on an assumption: Exclusion of patients registered only from a death certificate or at autopsy does not substantially bias the survival estimates.

Probably not a substantial claim: may not pass S3 (load-bearing).

Some centres in Nigeria and India changed hospital admission or pathology request forms to collect telephone numbers of patients and next of kin to aid follow-up.

“Some centres in Nigeria and India have modified hospital admission forms or pathology request forms to capture telephone numbers of patients and their next of kin, to facilitate follow-up of their patients.”

May combine two claims.

For most cancers, survival increased more in Denmark, and Danish survival has nearly caught up with the other Nordic countries.

“For most cancers, survival has increased more in Denmark, and survival has nearly caught up with the other Nordic countries.81,82”

• cites [81] Engholm, 2016Not checked yet

  Not checked in the free run.

• cites [82] Storm, 2010, Survival of patients diagnosed with cancer in the Nordic countries up to 1999-2003 followed to the end of 2006. A critical overview of the results, Acta Oncol sourceNot checked yet

  Not checked in the free run.

May combine two claims.

For ovarian cancer, age-standardised five-year net survival was mostly between 30% and 50%, with even wider variation in Europe and Asia.

“Age-standardised five-year net survival was mostly in the range 30–50%, with even wider variation in Europe and Asia (web-figure 4).”

May combine two claims.

Cancer survival trends are generally increasing, even for some more lethal cancers.

“Cancer survival trends are generally increasing, even for some of the more lethal cancers.”

May combine two claims.

Data quality improved in many countries, with more cases confirmed by pathology, imaging or biomarkers and fewer patients lost to follow-up.

“Data quality has improved in many countries, with increasing proportions of cases for which the diagnosis was confirmed by pathology, imaging or biomarkers, and reduction in the proportion of patients lost to follow-up.”

May combine two claims.

In Jordan, survival estimates for stomach and colorectal cancers were very high and were flagged as less reliable.

“Survival estimates were very high for stomach and colorectal cancers and they are flagged as less reliable.”

May combine two claims.

The overall proportion of tumour records excluded for incomplete dates was 0.5%, and for other reasons such as missing vital status was 1.2%; both remained very low.

“The overall proportion of tumour records excluded because of incomplete dates (0.5%), or for other reasons such as missing vital status (1.2%), has remained very low.”

May combine two claims.

For lymphoid malignancies, five-year age-standardised net survival was usually between 40% and 70% in most countries, lower in Asia and in Central and South America.

“Five-year age-standardised net survival was usually in the range 40–70% in most countries, lower in Asia and in Central and South America (web-figure 4).”

May combine two claims.

Five-year survival trends for childhood lymphoma were generally rather flat over the 15 years between 2000-2004 and 2010-2014, but survival increased by 5-10% in the US, Korea, Singapore, Taiwan, and four European countries.

“Five-year survival trends were generally rather flat over the 15 years between 2000–2004 and 2010–2014 (web-figure 3), but survival increased by 5–10% in the US; in Korea*, Singapore* and Taiwan*; and in 4 European countries (North: the UK*; South: Portugal*, Spain; West: Germany).”

Probably not a substantial claim: may not pass S3 (load-bearing).

Many registries reported that CONCORD data quality reports helped them improve their data.

“Many registries told us that the CONCORD data quality reports helped them improve their data.”

May combine two claims.

Trends in five-year survival for brain cancer between 2000-2004 and 2010-2014 were generally rather flat, but survival increased by 3-5% in 15 countries.

“Trends in 5-year survival between 2000–2004 and 2010–2014 were generally rather flat, but survival increased by 3–5% in 15 countries: Martinique* and Puerto Rico*; in Canada; in Israel* and Thailand; in 9 European countries (North: Iceland*, Latvia*, Lithuania*, Norway*, Sweden*; South: Croatia*, Italy; West: France, Switzerland), and in New Zealand*.”

May combine two claims.

The overall proportion of cancers registered only from a death certificate or detected at autopsy fell to 2.9%, but remains high in some countries with slow registration, especially for oesophageal, pancreatic and liver cancers.

“The overall proportion of cancers registered solely from a death certificate (DCO), or detected at autopsy, dropped to 2.9%, but it remains high in some countries where cancer registration processes are slow, especially for the more rapidly lethal cancers of the oesophagus, pancreas and liver.”

May combine two claims.

Pancreatic cancer survival trends between 2000-2004 and 2010-2014 were generally flat, but increases of 3-5% occurred in Canada, the US, Korea, Singapore, several European countries, and Australia.

“Trends in 5-year survival between 2000–2004 and 2010–2014 were generally flat, but increases of 3–5% were seen in Canada and the US; in Korea* and Singapore*, in 12 European countries (North: Denmark*, Estonia*, Ireland*, Latvia*, Norway*, Sweden*, the UK*; South: Portugal*; East: Czech Republic*; West: Belgium*, the Netherlands*, Switzerland), and in Australia* (web-figure 3).”

May combine two claims.

For myeloid malignancies, age-standardised five-year net survival was between 30% and 50% in most countries, although lower in Asia.

“Age-standardised five-year net survival was in the range 30–50% in most countries, although lower in Asia (web-figure 4).”

Probably not a substantial claim: may not pass S3 (load-bearing).

Some countries with incomplete death registration, such as India, actively follow up patients to determine vital status.

“Other countries with incomplete death registration, such as India, follow up their patients actively to determine their vital status.”

May combine two claims.

Age-standardised five-year net survival for stomach cancer was below 30% in most countries but high in Korea (69%) and Japan (60%), where it increased by up to 10% between 2000-2004 and 2010-2014.

“Age-standardised five-year net survival for stomach cancer was below 30% in most countries, but high in Korea (69%) and Japan (60%), where it increased by up to 10% between 2000–2004 and 2010–2014.”

May combine two claims.

Liver cancer survival trends were generally stable from 1995 to 2014, but survival increased by more than 10% in Korea, Singapore and Norway.

“For example, survival trends for liver cancer were generally stable during 1995 to 2014, but survival increased by more than 10% in Korea, Singapore and Norway.”

May combine two claims.

In most south-east Asian countries, survival for myeloid malignancies rose by 5-14%, and for lymphoid malignancies by 10% or more.

“In most south-east Asian countries, survival for myeloid malignancies has risen by 5–14%, and by 10% or more for lymphoid malignancies.”

May combine two claims.

For Nigeria (Ibadan) and South Africa (Eastern Cape), data were available for only 12 and 7 children, respectively, and survival was not estimated.

“For Nigeria (Ibadan) and South Africa (Eastern Cape), data were only available for 12 and 7 children, respectively, and survival was not estimated.”

May combine two claims.

In Canada, national coverage of cancer survival statistics was first achieved in CONCORD-2 with data from all 13 provinces and territories for 1995-2009, but several jurisdictions could not participate in CONCORD-3 because of legal or administrative difficulties linking their cancer registry with death records.

“In Canada, for example, national coverage of cancer survival statistics was achieved for the first time in CONCORD-2,6 with data from all 13 provinces and territories for 1995–2009, but several jurisdictions were unable to participate in CONCORD-3 because of legal or administrative difficulties in linking their cancer registry with death records.”

• cites [6] Allemani, 2015, Global surveillance of cancer survival 1995–2009: analysis of individual data for 25 676 887 patients from 279 population-based registries in 67 countries (CONCORD-2), Lancet sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

In some countries, obtaining the death and population counts needed to construct life tables has become more difficult.

“In some countries, it has become more difficult to obtain the death and population counts required to construct life tables.”

May combine two claims.

Rectal cancer survival was very similar among Nordic countries (64-65%) and most southern European countries (61%).

“Survival for rectal cancer was very similar among the Nordic countries (64–65%) and among most southern European countries (61%).”

May combine two claims.

Some registries still do not record in situ tumours, and others did not submit in situ tumour data.

“Some registries still do not record in situ tumours, while other registries did not submit data for in situ tumours.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Legal, administrative and technical difficulties in linking cancer registries to death indexes have been reported.

“Legal and administrative obstacles and technical difficulties have all been reported.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Cancer registries report increasing difficulty linking their databases to regional or national death indexes.

“Yet cancer registries report increasing difficulty in linking their databases to regional or national death indexes.”

May combine two claims.

Some registries do not record benign brain tumours, and this varies both between and within countries.

“However, some registries do not record benign brain tumours, and this varies both between and within countries.”

May combine two claims.

Survival from chronic lymphocytic leukaemia is much lower in Taiwan than in the US, and the CONCORD findings show generally lower adult survival in south-east Asia than in other countries for both myeloid and lymphoid malignancies.

“However, survival from CLL is also much lower in Taiwan than in the US,92 and the findings reported here also show that survival is generally lower in south-east Asia than in other countries in adults, for both myeloid and lymphoid malignancies.”

• cites [92] Wu, 2013, Improving but inferior survival in patients with chronic lymphocytic leukemia in Taiwan: a population-based study, 1990–2004, PLoS One sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Where survival could be estimated with some confidence in Africa, it was often very low, although survival in Mauritius was generally higher.

“Where survival could be estimated with some confidence, it was often very low, although survival in Mauritius was generally higher.”

May combine two claims.

The increasing trend in five-year net survival from breast cancer during 1995-2009 continued in most countries up to 2014, but survival remains lower in India, Thailand and several eastern European countries.

“The increasing trend in five-year net survival from breast cancer during the 15 years 1995–2009 has continued in most countries up to 2014, but it remains lower in India, Thailand and several of the eastern European countries.”

May combine two claims.

In CONCORD data, acral lentiginous melanoma represents 1.2% of all skin melanomas, 1% in Europe and North America, and 6% in Asia.

“In the CONCORD data (not shown), acral lentiginous melanoma represents 1.2% of all skin melanomas, and 1% in Europe and North America, but 6% in Asia.”

May combine two claims.

Benign brain tumours are not registrable in New South Wales or Western Australia, which together hold 45% of the national population, whereas they make up up to 5% of brain tumours in Queensland and Victoria, which have a similar combined population.

“For example, benign brain tumours are not registrable in New South Wales or Western Australia (45% of the national population), whereas they comprise up to 5% of brain tumours in Queensland and Victoria, with a similar combined population.”

May combine two claims.

Survival from brain tumours in children is generally higher than in adults, but the global range is much wider.

“Survival from brain tumours in children is generally higher than for adults, but the global range is much more pronounced.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Seven of the 329 registries that submitted data were excluded for protocol non-compliance that could not be corrected in time.

“Of the 329 registries that submitted data, 7 were excluded because their data were not compliant with the protocol and could not be rectified in time.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Around 2012, about 14 million people worldwide were diagnosed with cancer annually, and this was projected to rise by 50% to 21.6 million per year by 2030.

“an estimated 14 million patients a year world-wide around 2012,2 with a 50% projected increase to 21.6 million patients a year by 2030.3”

• cites [2]Not checked yet

  Not checked in the free run.

• cites [3] 2016Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Forty countries had complete national population coverage in CONCORD-2.

“In 40 countries, the data had 100% national population coverage.”

• cites [6] Allemani, 2015, Global surveillance of cancer survival 1995–2009: analysis of individual data for 25 676 887 patients from 279 population-based registries in 67 countries (CONCORD-2), Lancet sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The CONCORD-3 protocol was translated into eight additional languages and verified against the English version by other native speakers.

“the CONCORD-3 protocol was translated into eight other languages: Arabic, Chinese (Mandarin), French, Italian, Japanese, Portuguese, Russian and Spanish. Translations were done by native speakers in the CONCORD Central Analytic Team in London or the wider CONCORD Working Group, and checked against the English original by other native speakers.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

CONCORD-2 used central quality control and analysed individual records for 25,676,887 patients with one of 10 common cancers, which made up 63% of the 2009 global cancer burden.

“CONCORD-2 incorporated centralised quality control and analysis of individual data for 25,676,887 patients diagnosed with one of 10 common cancers that represented 63% of the global cancer burden in 2009.”

• cites [6] Allemani, 2015, Global surveillance of cancer survival 1995–2009: analysis of individual data for 25 676 887 patients from 279 population-based registries in 67 countries (CONCORD-2), Lancet sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The 18 index cancers made up roughly 75% of global cancer diagnoses around 2012 (10,537,411 of 14,067,894, excluding non-melanoma skin cancer).

“Collectively, these cancers accounted for about 75% of the estimated number of patients diagnosed with cancer world-wide each year around 2012 (10,537,411 of 14,067,894, excluding non-melanoma skin cancer; Table 1).13”

• cites [13] Weir, 2017, Population-based cancer survival (2001–2009) in the United States: findings from the CONCORD-2 study, Cancer sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

CONCORD-2's 279 registries together covered 896 million people.

“The 279 population-based registries covered a combined total population of 896 million people.”

• cites [6] Allemani, 2015, Global surveillance of cancer survival 1995–2009: analysis of individual data for 25 676 887 patients from 279 population-based registries in 67 countries (CONCORD-2), Lancet sourceNot checked yet

  Not checked in the free run.

May combine two claims.

CONCORD-2 reported survival trends for 1995-2009 in 67 countries whose combined population was two-thirds (4.8 billion) of the world total.

“with publication of trends in survival over the 15-year period 1995–2009 among patients diagnosed with cancer in 67 countries that were home to two-thirds (4.8 billion) of the world’s population.”

• cites [6] Allemani, 2015, Global surveillance of cancer survival 1995–2009: analysis of individual data for 25 676 887 patients from 279 population-based registries in 67 countries (CONCORD-2), Lancet sourceNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

Survival estimates were standardised by age using the International Cancer Survival Standard weights.

“Estimates were age-standardised with the International Cancer Survival Standard weights.”

Classified as probably not a substantial claim. May combine two claims.

Quality control followed standard procedures, and participating registries corrected errors.

“Standardised quality control procedures were applied; errors were rectified by the registry concerned.”

Probably not a substantial claim: may not pass S2 (checkable).

CONCORD adds evidence for global cancer-control policy.

“It contributes to the evidence base for global policy on cancer control.”

• AssumptionRests on an assumption: Population-based cancer survival can serve as a metric for evaluating access to effective treatment and health-system performance.

Probably not a substantial claim: may not pass S3 (load-bearing).

Patients recorded as having migrated beyond the registry's jurisdiction or to another country may be recorded as lost to follow-up and censored from survival analysis on that date.

“Patients recorded as having migrated beyond the registry’s jurisdiction, or to another country, may be recorded as lost to follow-up, because the patient’s eventual death is unlikely to be recorded: they are censored from survival analysis on that date.”

• AssumptionRests on an assumption: Patients lost to follow-up or censored alive have the same subsequent survival as patients with complete follow-up.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

If 10-49 patients were available, survival was estimated only for all ages combined; if 50 or more, survival estimation was attempted for each age group.

“If 10–49 patients were available for analysis in a given calendar period, we only estimated survival for all ages combined. If 50 or more patients were available, we attempted survival estimation for each age group.”

Classified as probably not a substantial claim. May combine two claims.

For children, survival was estimated for age groups 0-4, 5-9 and 10-14 years, and age-standardised estimates were obtained by assigning equal weights to the three age-specific estimates.

“For children, we estimated survival for the age groups 0–4, 5–9 and 10–14 years; age-standardised estimates were obtained by assigning equal weights to the three age-specific estimates.41,42”

• cites [41] Corazziari, 2004, Standard cancer patient population for age standardising survival ratios, Eur J Cancer sourceNot checked yet

  Not checked in the free run.

• cites [42] Stiller, 1990, Trends in survival for childhood cancer in Britain diagnosed 1971–85, Br J Cancer sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The study maintained ethics approvals from the UK Health Research Authority, NHS Research Ethics Service, and London School of Hygiene and Tropical Medicine, and statutory/ethical approvals and data sharing agreements in 85 other jurisdictions.

“We maintain approvals from the Ethics and Confidentiality Committee of the UK’s statutory Health Research Authority (HRA) (reference ECC 3–04(i)/2011, last update 3 March 2017) the National Health Service Research Ethics Service (11/LO/0331, 21 February 2017) and the London School of Hygiene and Tropical Medicine (12171, 6 September 2017).”

Classified as probably not a substantial claim.

Pooled estimates for countries with more than one registry did not include data from registries for which estimates were less reliable.

“The pooled estimates for countries with more than one registry do not include data from registries for which the estimates were less reliable.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Quality checks had three stages: protocol adherence, exclusions, and editorial review.

“we performed data quality checks in three phases: protocol adherence, exclusions, and editorial checks.”

Classified as probably not a substantial claim.

The cohort approach is considered the gold standard because it provides a survival estimate for patients diagnosed in the same period, likely treated similarly, and followed for at least the required duration.

“The cohort approach is considered the gold standard,29,30 because it provides a survival estimate for a group of patients who were diagnosed during the same year or period, are likely to have been treated in similar fashion, and who have all been followed up for at least the duration of survival required, such as 5 years.”

• cites [29] Allemani, 2017, Population-based cancer survival in the US: data, quality control and statistical methods, Cancer sourceNot checked yet

  Not checked in the free run.

• cites [30] Estève, 1994Not checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

Patients with in situ cancer were excluded from survival analyses.

“Patients with in situ cancer were not included in survival analyses.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Identity-number transcription errors can occur, so name, sex and birth date are often used to improve match accuracy.

“Transcription errors can arise with identity numbers, so variables such as the name, sex and date of birth are often used to improve the probability of an accurate match between a cancer record and a death registration.”

Classified as probably not a substantial claim.

Patients registered solely from a death certificate or diagnosed at autopsy were excluded from analyses because their survival time is unknown.

“Patients registered solely from a death certificate or diagnosed at autopsy were excluded from analyses, because their survival time is unknown.”

• AssumptionRests on an assumption: Exclusion of patients registered only from a death certificate or at autopsy does not substantially bias the survival estimates.

Classified as probably not a substantial claim. May combine two claims.

For adults, ICSS weights were used with age at diagnosis categorised into five groups, and group 2 weights were used for melanoma, cervix and brain, group 1 for oesophagus, stomach, colon, rectum, liver, pancreas, lung, breast, ovary and prostate, and both groups for haematopoietic malignancies.

“For adults, we used the International Cancer Survival Standard (ICSS) weights,41 in which age at diagnosis is categorised into 5 groups: 15–44, 45–54, 55–64, 65–74 and 75–99 years and, for prostate cancer, 15–54, 55–64, 65–74, 75–84 and 85–99 years. Of the three sets of ICSS weights, we used group 2 (cancers for which incidence does not increase steeply with age) for melanoma of the skin, cervix”

• cites [41] Corazziari, 2004, Standard cancer patient population for age standardising survival ratios, Eur J Cancer sourceNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim. May combine two claims.

We checked morphological verification rates, non-specific morphology rates, date distributions, and proportions of patients who died within 30 days, were lost to follow-up, or were censored within five years.

“Finally, we examined the proportion of tumour records with morphological verification of the diagnosis, whether from histology of a biopsy or surgical specimen, cytology of a smear or bone marrow aspirate, or from imaging or biomarkers, including tumours with a specific morphology code. We also examined the proportion of cases with non-specific morphology, the distributions of the day and month of”

Classified as probably not a substantial claim. May combine two claims.

Survival estimates were flagged as less reliable if 15% or more of patients were lost to follow-up or censored alive within five years of diagnosis, or before 31 December 2014 for patients diagnosed in 2010 or later.

“Survival estimates that are considered less reliable are flagged (§) in tables and graphics. These are data sets in which 15% or more of patients were lost to follow-up or censored alive within five years of diagnosis. For patients diagnosed in 2010 or later, this criterion was applied for patients censored alive before 31 December 2014, the study closure date.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Active follow-up techniques were used by 124 (51.0%) registries, which routinely contact the treating physician, GP or hospital administration to determine vital status.

“Active follow-up techniques are also used by 124 (51.0%) registries, which routinely contact the treating physician, GP or hospital administration to determine the vital status for each registered patient, often on a quarterly or annual basis.”

Classified as probably not a substantial claim. May combine two claims.

For each index cancer and period, registries received exclusion reports on error types, DCO/autopsy cases, and eligible patients included; they could request details, and many updated their databases.

“Registries were sent ‘exclusion reports’ for each index cancer and each calendar period, summarising the number of tumour records with each type of definite or possible error, the number registered from a death certificate only (DCO) or detected at autopsy, and the number and proportion of eligible patients whose data could be included in survival analyses. Registries were invited to request”

Probably not a substantial claim: may not pass S3 (load-bearing).

Less reliable estimates are reported with a flag when they are the only available information from a given country or territory.

“Less reliable estimates are reported with a flag when they are the only available information from a given country or territory.”

Probably not a substantial claim: may not pass S3 (load-bearing).

A further 116,316 (0.3%) patients were excluded for other reasons including definite errors and possible inconsistencies.

“and 116,316 (0.3%) for other reasons. These included definite errors (unknown vital status, unknown sex, sex-site error, invalid date(s) or sequence of dates) and possible errors, such as apparent inconsistencies between age, cancer site and morphology (details on request).”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Data were transmitted via a specially configured file transmission utility with 256-bit Advanced Encryption Security, using a random one-time password for each data file.

“Participants transmit data via a specially configured file transmission utility with 256-bit Advanced Encryption Security. The utility automatically generates a random, strong, one-time password for each data file at the time of transmission, and emails it to a different address.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The funding sources played no part in the design, data collection, quality control, analysis, interpretation, writing, or decision to submit for publication.

“The funding sources played no part in the design, data collection, quality control, analysis, interpretation of the findings, writing of the manuscript, or the decision to submit for publication.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Survival estimates were also flagged as less reliable if 15% or more of patients were registered only from a death certificate or at autopsy and excluded from analysis.

“Estimates are also flagged as less reliable if 15% or more of patients were registered only from a death certificate or at autopsy and excluded from analysis, since their survival is unknown.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Survival was not estimated if fewer than 10 patients were available for analysis.

“Survival was not estimated if fewer than 10 patients were available for analysis.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Exclusion criteria were set in advance from CONCORD-2 experience and expanded for some new cancers and 14 extra variables on grade and treatment.

“Exclusion criteria were defined a priori, based on experience from CONCORD-2, and extended to cover features of some of the five additional cancers such as Ann Arbor stage for the lymphomas, and 14 additional variables on tumour grade and treatment.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Survival estimates were also flagged if 15% or more of patients were excluded because one or more dates was incomplete.

“Finally, estimates are also flagged if 15% or more of patients were excluded from analysis because one or more dates was incomplete: unknown year of birth, unknown month or year of diagnosis, or unknown year of last vital status.”

Classified as probably not a substantial claim.

Age-standardised net survival estimates at five years after diagnosis were presented for each country, registry and calendar period.

“For each country, registry and calendar period, we present age-standardised net survival estimates for each cancer at five years after diagnosis.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Some registries routinely get death certificates, but this alone is insufficient because certificates not mentioning cancer are rarely included.

“Some registries routinely receive paper or electronic death certificates for their territory, but this is insufficient on its own, since death certificates that do not mention cancer are rarely included.”

Classified as probably not a substantial claim.

NAACCR updated a CONCORD-2 program so all North American registries could recode their full incidence databases to IACR multiple-primary rules before extracting 2000-2014 data for CONCORD-3.

“The North American Association of Central Cancer Registries (NAACCR) kindly updated the program developed for CONCORD-2 to enable all North American registries to recode their entire incidence databases to the IACR multiple primary rules, before their data sets for 2000–2014 were extracted for CONCORD-3.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Each registry received a report on protocol compliance for 51 variables per cancer file; any compliance below 100% required correction or resubmission.

“First, registries were sent a report showing the percentage compliance with the protocol for each of 51 variables in each cancer file. Compliance less than 100% required correction and/or re-submission of data.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Tumour records were effectively anonymised: they did not contain patient name, address, postcode, or national identity or social security number, and all variables were numeric or alphanumeric codes.

“Tumour records are effectively anonymised: they do not contain the patient’s name, address, postcode, or any national identity or social security number. All variables are numeric or alphanumeric codes.”

Classified as probably not a substantial claim. May combine two claims.

Registries were asked to submit data with follow-up for at least 5 years or, for patients diagnosed during 2010-2014, until 31 December 2014.

“Registries were asked to submit data with follow-up for at least 5 years or, for patients diagnosed during 2010–2014, until 31 December 2014.”

May combine two claims.

The cohort approach was used for patients diagnosed in 2000-2004 and 2005-2009 because in most data sets all patients had been followed up for at least five years.

“We used the cohort approach for patients diagnosed in 2000–2004 and 2005–2009, because in most data sets, all patients had been followed up for at least five years.”

Classified as probably not a substantial claim.

For 64 registries that provided abridged or unsmoothed complete mortality rates, the Ewbank relational model with 3 or 4 parameters was used to interpolate and smooth mortality rates against a high-quality smooth life table from a similar country.

“For 64 registries that provided abridged mortality rates, or complete mortality rates that were not smoothed, we used the Ewbank relational model38 with 3 or 4 parameters to interpolate (if abridged) and smooth the mortality rates for the registry territory against a high-quality smooth life table for a country with a similar pattern of mortality by age.39”

• cites [38] Ewbank, 1983, A reducible four-parameter system of model life tables, Popul Studies sourceNot checked yet

  Not checked in the free run.

• cites [39] Brass, 1971, On the scale of mortalityNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Clinicians have used five-year survival as a measure of cancer treatment effectiveness for over six decades.

“Five-year survival has been recognised by clinicians as an index of the effectiveness of the treatment of cancer for more than 60 years.”

• cites [44] Berkson, 1952, Survival curve for cancer patients following treatment, J Am Stat Assoc sourceNot checked yet

  Not checked in the free run.

• cites [45] Dickman, 2006, Interpreting trends in cancer patient survival, J Intern Med sourceNot checked yet

  Not checked in the free run.

May combine two claims.

The OECD's inclusion of CONCORD estimates formally recognises their global coverage, methodological rigour and international comparability, and they will contribute to comparing health system performance in 48 countries, including all OECD members.

“This is formal recognition of the global coverage, methodological rigour and international comparability of the CONCORD survival estimates, which will now contribute to the comparative evaluation of health systems performance in 48 countries, including all OECD Member countries.”

Probably not a substantial claim: may not pass S2 (checkable).

Population-based cancer survival is increasingly viewed as a key indicator of how well health systems manage cancer care and treatment overall.

“Population-based cancer survival is increasingly recognised as a key indicator of the overall effectiveness of the health systems in managing care and treatment for all cancer patients.46,47”

• cites [46] Coleman, 2014, Cancer survival: global surveillance will stimulate health policy and improve equity, Lancet sourceNot checked yet

  Not checked in the free run.

• cites [47] Allemani C. The importance of global surveillance of cancer survival for cancer control: the CONCORD programme. Cancer Control (in press).Not checked yet

  Not checked in the free run.

May combine two claims.

The study provides internationally comparable survival trends for 18 cancers, which together account for 75% of all cancers diagnosed globally each year.

“Internationally comparable survival trends are now available for 18 cancers that collectively represent 75% of all cancers diagnosed world-wide every year.”

• AssumptionRests on an assumption: The population-based cancer registries record all cancers diagnosed in their coverage areas.
• AssumptionRests on an assumption: The participating countries' data are representative of the world's population for the purpose of global surveillance.
• AssumptionRests on an assumption: The survival estimates from different countries are internationally comparable.

May combine two claims.

Most lung cancer survival estimates in Central and South America were less reliable due to the high proportion of DCO registrations excluded from analysis, although the proportion of DCOs has generally decreased world-wide.

“Most estimates in Central and South America were less reliable, due to the high proportion of DCO registrations excluded from analysis (Table 5, web-table 4), although the proportion of DCOs has generally decreased world-wide (Table 3, web-table 3).”

Probably not a substantial claim: may not pass S1 (asserted), S3 (load-bearing), S4 (based).

Most myeloid malignancy survival estimates were considered reliable.

“Most estimates were considered reliable (Table 5, web-table 5).”

Classified as probably not a substantial claim. May combine two claims.

The authors performed extensive data quality checks and worked with cancer registries to resolve problems.

“We carried out extensive checks on data quality, and liaised with the cancer registries to resolve problems.”

May combine two claims.

The quality and completeness of cancer registration data and follow-up differ across countries and can affect how comparable survival estimates are.

“The quality and completeness of cancer registration data and follow-up vary between countries, and this can affect the comparability of survival estimates.”

May combine two claims.

CONCORD-3 adopted a design and statistical approach comparable to CONCORD-2, allowing assessment of survival trends for 10 cancers across 1995-2014.

“We used a similar design and statistical methods to those used in CONCORD-26, to enable evaluation of survival trends for 10 cancers over the 20-year period 1995–2014.”

• cites CONCORD-2Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable).

Cancer survival data are useful for cancer control and health policy at state, national and global levels in both high-income and low-income countries.

“Cancer survival has applications to cancer control and health policy at the state, national and global levels, in both high-income and low-income countries.62”

• cites [62] Coleman, 2013, Why we need global surveillance of cancer survival: the CONCORD programmeNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim. May combine two claims.

The study is not a compilation of published reports or a meta-analysis.

“This is not a compilation of published reports or a meta-analysis.”

May combine two claims.

The proportion with pathological confirmation varies widely by country and cancer, but routine registry evidence for a cancer diagnosis is much more precise and definitive than the causes of death recorded on death certificates used for mortality statistics.

“The proportion varies widely between countries and for different cancers, but the evidence supporting a cancer diagnosis in routine cancer registry data is far more precise and definitive than for the cause(s) of death recorded on death certificates, from which mortality statistics are obtained.72–75”

• cites [72] Laurenti, 2000, Accuracy of statements of the cause of death on death certificates and the international comparability of mortality statisticsNot checked yet

  Not checked in the free run.

• cites [73] Ovcharov, 2000, The improvement of statistical records of the causes of death in RussiaNot checked yet

  Not checked in the free run.

• cites [74] Filippatos, 2016, The quality of death certification practice in Greece, HippokratiaNot checked yet

  Not checked in the free run.

• cites [75] Haghighi, 2014, Impact of documentation errors on accuracy of cause of death coding in an educational hospital in Southern Iran, HIM J sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Death-certificate-only cases can be included in cancer incidence statistics under certain assumptions, but they reflect some underestimation of incidence.

“DCOs can be included in cancer incidence statistics under certain assumptions, but they reflect some under-estimation of incidence.69”

• cites [69] 1991Not checked yet

  Not checked in the free run.

May combine two claims.

In some settings, especially South America, improved data quality is reflected in lower survival estimates than previously published, with fewer death-certificate-only registrations and more complete follow-up.

“In some cases, especially in South America, the improvement in data quality is reflected in survival estimates that are actually lower than those previously published,6 with fewer death-certificate-only registrations and more complete follow-up.”

• cites [6] Allemani, 2015, Global surveillance of cancer survival 1995–2009: analysis of individual data for 25 676 887 patients from 279 population-based registries in 67 countries (CONCORD-2), Lancet sourceNot checked yet

  Not checked in the free run.

May combine two claims.

In some countries, survival estimates fluctuated or declined across calendar periods, likely because completeness of cancer registration and vital-status follow-up improved.

“In some countries, survival estimates have fluctuated or declined in successive calendar periods; this is likely to reflect improvements in the completeness of cancer registration data and in the completeness of follow-up for vital status.”

Probably not a substantial claim: may not pass S2 (checkable).

Cancer survival estimates are needed to assess the overall effectiveness of a country's health system.

“On the contrary, cancer survival estimates are required to assess the overall effectiveness of a country’s health system in the first place.”

Probably not a substantial claim: may not pass S2 (checkable).

Cancer registries are essential for assessing the overall effectiveness of a country's health system.

“For that, cancer registries are essential.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The authors provided extensive documentation of data quality using standard indicators for each cancer and each cancer registry.

“We have provided extensive documentation of data quality with standard indicators69 for each cancer and each cancer registry (web-table 2).”

• cites [69] 1991Not checked yet

  Not checked in the free run.

May combine two claims.

The reported survival estimates come directly from individual patient records and long-term vital-status follow-up, followed by standardised quality control and central analysis.

“The survival estimates reported here are derived directly from the records of individual patients diagnosed with cancer, and from long-term follow-up to ascertain their vital status, followed by standardised quality control and central analysis.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The OECD concluded from wide international variation in cancer survival that many countries could improve their cancer control.

“The Organisation for Economic Co-operation and Development (OECD) recently concluded from the wide international variation in cancer survival that many countries could do better in cancer control.64”

• cites [64] 2013Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Five-year survival has risen for many cancers since the 1950s, yet it is still a common benchmark and does not directly equal the share of patients cured.

“Five-year survival has increased for many cancers since the 1950s, but it remains a widely used benchmark, even though it cannot be directly interpreted as the proportion of patients who are cured.45”

• cites [45] Dickman, 2006, Interpreting trends in cancer patient survival, J Intern Med sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S1 (asserted), S3 (load-bearing), S4 (based).

Most breast cancer survival estimates were considered reliable.

“Most estimates were considered reliable (Table 5, web-table 4).”

Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

Population-based survival reflects the overall effectiveness of a health service and depends on factors broader than the competence of any individual clinician or team.

“Population-based survival reflects the overall effectiveness of the health service, which depends on much wider issues than the competence of any individual doctor or team.64–66”

• cites [64] 2013Not checked yet

  Not checked in the free run.

• cites [65] Atun, 2015, Expanding global access to radiotherapy, Lancet Oncol sourceNot checked yet

  Not checked in the free run.

• cites [66] Meara, 2015, Global Surgery 2030: evidence and solutions for achieving health, welfare, and economic development, Lancet sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

A recent hospital-based study from Jordan suggests that colon cancer survival there is much lower than the CONCORD estimates.

“One recent hospital-based survival study from Jordan71 suggests that colon cancer survival in Jordan is much lower than the estimates we have obtained.”

• cites [71] Abdel-Razeq, 2015, Cancer care in Jordan, Hematol Oncol Stem Cell Ther sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S1 (asserted), S3 (load-bearing), S4 (based).

Most lymphoid malignancy survival estimates were considered reliable.

“Most estimates were considered reliable (Table 5, web-table 5).”

Probably not a substantial claim: may not pass S3 (load-bearing).

The authors extended the data quality control programs and reports.

“We extended the data quality control programs and the reports.”

May combine two claims.

Death-certificate-only cases must be excluded from survival analyses because the patient's survival time is unknown, and including them would tend to inflate survival estimates.

“By contrast, DCO cases must be excluded from survival analyses, because the patient’s survival time is unknown. This tends to inflate survival estimates.70”

• cites [70] Berrino, 1995, Basic issues in the estimation and comparison of cancer patient survivalNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

To control for background mortality by age and sex, the authors updated the life-table library for 1995-2010 by country, registry, race in selected countries, and calendar year to 2014, with a statistical summary for each set.

“To control for background mortality by age and sex, we updated the library of life tables for 1995–201097 by country, registry, race (selected countries) and calendar year to 2014, with a statistical summary for each set of life tables.”

• cites [97] Ginsburg, 2017, The global burden of women's cancers: a grand challenge in global health, Lancet sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The global economic cost of treating the 12.9 million new cancer patients diagnosed worldwide in 2009 was estimated at US$286 billion.

“The global economic cost of treating the 12.9 million new cancer patients diagnosed worldwide in 2009 was estimated at US$286 billion.100”

• cites [100] Mariotto, 2011, Projections of the cost of cancer care in the United States: 2010–2020, J Natl Cancer Inst sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Civil registration and vital statistics systems support basic government functions and enable measurement of progress toward development goals, but population coverage is poor in low- and middle-income countries and closely related to gross national income.

“These systems support the basic functions of government and enable measurement of progress towards development goals,109 yet population coverage is poor in low- and middle-income countries, and closely related to gross national income.110”

• cites [109] Setel, 2007, Who counts? 1: A scandal of invisibility: making everyone count by counting everyone, Lancet sourceNot checked yet

  Not checked in the free run.

• cites [110] 2017Not checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

Passive follow-up requires efficient civil registration systems that capture information on all deaths.

“For passive follow-up to work, efficient civil registration systems that capture information on all deaths are required.”

Probably not a substantial claim: may not pass S2 (checkable).

Population-based cancer survival trend data that are comparable within and between countries are part of the evidence base needed to drive a policy shift toward equitable access to affordable cancer treatment.

“Population-based data on cancer survival trends that are comparable within and between countries are part of the evidence base needed to drive such a policy shift.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Where universal health coverage has not been achieved, out-of-pocket cancer care costs can cause financial catastrophe for individuals and families.

“Where universal health coverage has not been achieved, the out-of-pocket costs can lead to financial catastrophe for individuals and families.3”

• cites [3] 2016Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Rising cancer treatment costs threaten the viability of health systems and national economies.

“Spiralling costs102 threaten the viability of health systems and national economies.”

• cites [102] Knaul, 2011Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Based on these figures, the global cost of cancer treatment and care in 2017 was plausibly already substantially above US$300 billion per year.

“On the basis of these figures, it seems plausible that the global cost of cancer treatment and care in 2017 must already be substantially higher than US$300 billion a year.”

Probably not a substantial claim: may not pass S3 (load-bearing).

US cancer treatment and care costs were projected to rise by 23% between 2010 and 2020, from US$128 billion to US$158 billion, based only on demographic change and with fixed incidence, survival and treatment costs.

“The costs of cancer treatment and care in the US alone were projected to rise by 23% between 2010 and 2020, from US$128 to US$158 billion, solely on the basis of demographic change, and with fixed incidence rates, survival probabilities and treatment costs.101”

• cites [101] Fojo, 2009, How much is life worth: cetuximab, non-small cell lung cancer, and the $440 billion question, J Natl Cancer Inst sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

International comparisons of pancreatic cancer survival include both ductal adenocarcinomas and the less common neuroendocrine tumours, which generally have higher survival.

“International comparisons of survival for pancreatic cancer include both ductal adenocarcinomas and the less common neuroendocrine tumours, for which survival is generally higher.”

Probably not a substantial claim: may not pass S3 (load-bearing).