Patterns and trends of hospital mortality due to noncommunicable diseases and injuries in Tanzania: a 10- year retrospective analysis

Background: Globally, non-communicable diseases (NCD) kill about 40 million people annually, with about three-quarters of the deaths occurring in low and middle-income countries. This study was carried out to determine the patterns, trends, and causes of non-communicable disease mortality in hospitals of Tanzania from 2006-2015. Methods: This retrospective study involved primary, secondary, tertiary, and specialized hospitals in Tanzania. Death statistics were extracted from inpatient department registers, death registers, and International Classication of Diseases (ICD) report forms. Variables collected were deceased’s age, sex, cause and date of death. The ICD-10 coding system was used to assign each death to its underlying cause. Data were analysed using STATA version 14. Results: A total of 247,976 deaths were reported during the 10 years (2006–2015) in 39 hospitals. Of the total deaths, 67,711 (27.3%) were due to non-communicable diseases (NCD) and injuries. Cardio-circulatory diseases (31.9%), cancers (18.6%), chronic respiratory diseases (18.4%), and injuries (17.9%) accounted for the largest proportion (86.8%) of deaths due to NCDs. The majority (57.6%) of deaths due to NCD occurred among males. Overall, the total deaths from NCDs increased by 153.3% from 4,298 in 2006 to 10,886 in 2015. The age group 15-59 years (53.4%) was the most affected category. The overall 10-year annual age-standardized mortality rate (ASMR) for all NCDs and injuries was 235.2 per 100,0000 population. It was higher for males (287.2/100,000) than for females (186.7/100,000). The annual ASMR increased from 11.9 in 2006 to 36.5 per 100,000 populations in 2015. Most of the NCD deaths occurred in the secondary (40.0%) and primary level hospitals (27.8%). There were variations in the type of non-communicable diseases by geographical distributions. Conclusions: There was a substantial in ASMR due to

Evidence indicates an increase in the burden of non-communicable diseases in Sub-Saharan Africa [9][10][11][12][13][14][15][16][17]; though speci c NCDs and their risk factors vary considerably between countries and between urban and rural settings [17]. Studies in Nigeria, Sudan, and Tanzania have reported that NCDs account for 81% of admissions among > 60 years old individuals [18]. In East Africa, 40% of deaths are attributable to noncommunicable diseases, which are expected to overtake communicable diseases as the leading causes of death over the next 20 years [19]. The rate of urbanization and unhealthy lifestyles associated with poverty and inequality has been described as the reasons for an increasing burden of NCD in Sub-Saharan Africa [17,20].
According to estimates by the World Health Organization, NCDs account for 33% of all deaths in Tanzania [21]. Several studies in Tanzania provide data for speci c geographical areas on the prevalence of overweight, diabetes, cardiovascular disease, and cancers [22][23][24][25][26][27][28]. A very high prevalence of elevated blood pressure among adults has been reported in Tanzania in recent years [23,26,27,29]. High prevalence rates of diabetes and chronic respiratory diseases have also been reported in the country [30,31]. Cases and deaths due to road tra c injuries have increased by 44% and deaths by 64%, respectively between 1990 and 2000 [32]. Most of the major NCDs share risk factors. Tobacco use, excessive alcohol consumption, poor diet and lack of physical activities have been described as the most common behavioural risk factors associated with the major NCDs in Tanzania [31]. These factors are closely correlated with social determinants such as education, income, occupation, gender and ethnicity [33].
Although several studies have addressed the prevalence and risk factors of NCDs in Tanzania, there is limited information on the causes and mortality patterns of in-hospital NCDs. In Tanzania, most often community-based data have been used as sources of mortality data [34][35][36][37][38][39]. Nonetheless, such data are only available every 5-10 years. In-hospital mortality records are likely to be another important source of information on deaths. The cause of deaths certi ed by a medical practitioner is considered as the gold standard for cause-of-death reporting [40,41]. In-hospital mortality statistics derived from information reported on death certi cates is among the most widely used and reliable sources of health data. However, only about one-third of countries in Sub-Saharan Africa collect in-hospital deaths where the underlying cause of death has been coded according to the International Classi cation of Diseases [42]. This study was, therefore, carried out to determine the patterns and trends of hospital mortality due to noncommunicable diseases and injuries in Tanzania from 2006 to 2015.

Study design and source of data
This retrospective study carried out from July-December 2016 involved 39 hospitals in Tanzania. The health system in Tanzania is structured into three functional levels: Primary level with district hospitals, secondary level with the regional referral hospitals, and tertiary level with zonal referral hospitals and the national hospital. There are also specialized hospitals that do not t directly into the administrative level. The hospitals included in this study at the tertiary level were Muhimbili National Hospital, Bugando Medical Centre, Kilimanjaro Christian Medical Centre, and Mbeya Zonal Referral Hospital. The specialized hospitals Mortality data were extracted from the inpatient department (IPD) registers, death registers and International Classi cation of Diseases (ICD)-10 report forms. Hospital records were extracted manually from sources and lled in a customized paper-based collection tool. The extraction process was done iteratively moving from one source to another until all sources were assessed and all death events related to NCDs and injuries that occurred in the hospital were collected. Variables collected were deceased's age, sex, cause and date of death. Details in data extraction have been described elsewhere [43].

Statistical analysis
The ICD-10 coding system was used to assign each death to its underlying cause. Data was entered and processed in Epi-Data Software (Odense Denmark, EpiData Association, 2010). Data were analysed using STATA version 14 (STATA Corp LLC, Texas, USA). The age-standardized mortality rate was calculated for each of the major NCDs. We examined for any signi cant 10-year changes in trends of mortality rate for each disease, hospital type, and geographical zone.

During the 10 years (2006-2015) NCDs and injuries accounted for about one-third of all in-hospital deaths
in Tanzania. The majority of deaths due to NCDs occurred among those 15-59 years old. Similar to our ndings, a study in Ethiopia reported that NCDs contribute to high proportions of mortality in hospitals [44].
The ndings of this study indicate that most of the deaths due to NCD and injuries affected the young adult group and males. Similar ndings have been reported elsewhere [45]. It has been described that as age increases one becomes more exposed to the risk factors for long periods until the complications develop and hence experiences the clinical syndromes of NCDs [46]. Like in our study, males accounted for most NCD deaths in studies in Brazil, Sri Lanka, and China [47][48][49]. Despite the observations that NCD deaths affected more males than females, several studies in Tanzania have reported that women have a higher risk of being obese than men, hence more vulnerable to several NCDs [50][51][52]. Overweight and obesity are major risk factors for diabetes, cardio-circulatory diseases, and cancer [54]. A recent systematic review has identi ed several factors related to gender differences in the burden of NCDs. These include gender roles, physical access to recreational facilities, and preferences for walking and engaging in physical activity [55].
Cardio-circulatory diseases were the number one leading NCD cause of deaths followed by cancer and chronic respiratory diseases. Like elsewhere in Sub-Saharan Africa, the ndings of this study indicate that the largest proportion of the NCD mortalities are due to cardio-circulatory diseases, cancers, chronic respiratory diseases, injuries, and diabetes [17,21,56,57]. Both cardio-circulatory diseases, cancers, and chronic respiratory diseases have also been described as among the most prevalent NCDs in Tanzania [58].
According to WHO estimates, cardio-circulatory diseases are responsible for 13% of the total deaths in Tanzania [5,21]. There has been an increased prevalence of cardio-circulatory death rates from 9-13% between 2012 and 2016 in Tanzania [5,59]. Like in this study, previous studies in Tanzania have reported cardio-circulatory diseases as among the major causes of deaths due to NCDs [60].
The ndings from this study have shown that both injuries have increased by about 63% during 10 years. The majority of the injuries in this study were associated with road tra c accidents, drowning, and animal bites. Road tra c injuries (RTI) have been reported to account for the largest proportion of unintentional injuries in low-and middle-income countries [61]. The overall 31.98/100,000 age-standardized mortality rate due to injuries in this study was similar to that reported in a population-based study in a rural district of Tanzania [39]. It has already been reported that more males than females are likely to die from injuries than women [39]. A previous descriptive analysis of road tra c injuries in Tanzania has shown that between 1990 and 2000 the number of RTI rose by 44% and that of death by 64% [32]. Globally, each year 5.1 million people die from injuries and a quarter of these deaths are due to road tra c accidents [5] and 90% occur in low-and middle-income countries [62].
An increasing trend in ASMR due to cancers and diabetes was observed throughout the 10 years. Recent estimates in Tanzania indicate that age-standardized in-hospital mortality rate for cancer is 47.8/ 100,000 population; with cervical, oesophageal, and liver cancers being among the top three causes of deaths in Tanzania [63]. Cancer has been described as a major emerging public health problem in Sub-Saharan Africa because of population aging and growth, as well as increased prevalence of key risk factors, including those associated with socio-economic transition [64]. Diabetes was also among the ve leading causes of deaths due to NCDs in Tanzania. World Health Organization's recent statistics indicate diabetes mellitus death in Tanzania accounts for 1.86% of total death and age-adjusted death rate is 30.24 per 100,000 populations [8] which is higher than the ndings from this study (17.21 per 100,000). As in our study where about 50% of the deaths due to diabetes affect the 15-49 years' age group, about threequarters of diabetes-related deaths occur in economically-productive persons under the age of 6o years [8].
Except for diabetes, the contributions of NCDs to the total deaths varied by geographical zones of the country and levels of hospitals. Chronic respiratory diseases accounted for about one-third of deaths in the Southern Western, Southern Highlands, and Central Zones. About half of all deaths due to cardiocirculatory diseases were reported from hospitals in the Southern, Lake Victoria, Western and Central zones. Eastern, Lake Victoria, and northern zones reported the largest proportion of deaths due to cancers.
Injuries accounted for the largest proportions of deaths in the Western, Central, and Southern Highlands. These variations are likely to indicate regional and demographic differences in NCD prevalence underlying differences in lifestyle, socioeconomic status, and access to healthcare [65].
This study is likely to have some limitations. We analysed data on broadly de ned disease categories such as chronic respiratory diseases, cardio-circulatory diseases, cancers, injuries, diabetes, brain disorders, and kidney diseases. The analysis could not make inferences on speci c diseases. Moreover, despite the usefulness of the information that is available from this study, the hospital data are likely to provide an incomplete picture of the burden of NCDs in Tanzania, as the majority of deaths occur in the community and their causes are not registered. Despite these limitations, this study consolidates information on the NCD mortality and highlights patterns and trends of major NCDs in Tanzania.

Conclusions
NCD and injuries account for about one-third of all in-hospital deaths in Tanzania. In-hospital mortalities due to NCDs in Tanzania have increased by over half during the 10 years. Cardio circulatory diseases, cancers and chronic respiratory diseases and are responsible for over two-thirds of the deaths due to NCD and the majority of deaths affect males. The young adult age category accounts for the majority of those dying from NCDs. There were regional differences in terms of the number of deaths and types of NCDs. If the current trends continue, the probability of dying prematurely from the four main NCDs is likely to increase in Tanzania. It is important that the government of Tanzania invest in better management of NCD which includes early detection, screening, and timely treatment at all levels of the health system. and Children and the President's O ce Regional Administration and Local Government through the respective Regional Administrative Secretaries and Hospital Authorities. No consent to participate was required due to the nature of the study which involved extraction of data from registers and report forms.

Consent for publication
Not applicable

Competing interests
The authors declare that they have no competing interests

Availability of data
The data that support the ndings of this study are available from the National Institute for Medical Research but restrictions apply to the availability of these data. Data are however available from the authors upon reasonable request and with permission of the Tanzania Medical Research Coordinating Committee.

Competing interests
None declared.

Funding
This study received nancial assistance from the Global Funds for AIDS, Tuberculosis and Malaria through the Tanzania Ministry of Health, Community Development, Gender, Elderly and Children. The funding agent had no role in the design of the study, the collection, analysis, and interpretation of data and writing the manuscript.
Authors' contributions LEGM, SFR designed the study. LEGM, CK, SFR, MGC collected the data. CK, SFR, LEGM, EK analysed and interpreted the data. LEGM drafted the manuscript. SFR, CK, MGC critically revised the manuscript for important intellectual contents. All authors have read and approved the manuscript, have full access to all of the data, and take responsibility for the integrity of the data and the accuracy of the data analysis.
Cause of death  Geographical distribution of major NCD mortality (Key: CRD= chronic respiratory diseases; CCD= cardiocirculatory diseases)