Demographic, health services and socio-economic factors associated with pulmonary tuberculosis mortality in Los Altos Region of Chiapas, Mexico

Abstract

Background Chiapas is one of the Mexican states having the highest rates of Pulmonary Tuberculosis (PTB), due to the numerous factors impeding its management and control (poverty, poor housing and nutrition, shortage of health resources, among others).

Objectives To analyse the PTB mortality of a cohort of patients in Los Altos Region of Chiapas, who had been diagnosed with PTB from January 1, 1998 to December 31, 2002; and, to identify demographic, socioeconomic and health services utilization factors, associated with death from PTB.

Methods Analysis of a cohort of patients aged over 14 years diagnosed with PTB in the above mentioned period (n = 431) in Los Altos region of Chiapas. The records of the Tuberculosis Programme were reviewed, and patients were located through a search attempting to locate them in their homes. Those found alive were interviewed and asked to provide sputum samples. In the case of deceased patients, a verbal autopsy was obtained from a member of their family.

Results The records of the PTB Programme in the area were incomplete and erroneous in many cases. The results of the home follow-up visits were: 208 (48%) patients located alive, five of whom were still PTB positive (three with multi-drug resistance); 145 (34%) could not be located and 78 (18%) had already died. Apparently, in at least 40 cases, the deaths were associated with PTB. Of these forty, 33 (83%) died without having received any medical care. The factors associated with dying from PTB were: 45 and over years of age (OR = 1.3; 95% CI = 0.98–1.3), 0–3 schooling years (OR = 3.3; 95% CI = 1.1–9.6), engaged in agriculture (OR = 2.2; 95% CI = 1.1–4.4), not living in main villages of their municipality (OR = 1.2; 95% CI = 1.0–1.3), living in a rural community (OR = 2.7; 95% CI = 1.1–6.8), not having been treated in DOTS (OR = 1.2; 95% CI = 1.0–1.3) and having defaulted from treatment (OR = 11.5; 95% CI = 5.3–24.8).

Conclusions The high rate of mortality due to PTB observed constitutes a serious public health problem deserving attention. There is an urgent need to evaluate and restructure the Tuberculosis Programme in the studied area.

Mexico, pulmonary tuberculosis, DOTS, mortality, poverty

Introduction

The Directly Observed Therapy Strategy (DOTS) is the best therapeutic strategy for containing Tuberculosis (TB) in the world.^1,2 Despite this, in many countries it has not been put fully into practice.³

According to official statistics, in 2005, Mexico achieved 100% coverage of DOTS in those municipalities considered high priority.⁴ However, within the country there are differences in the instrumentalization of TB control programmes. Although officially, the state of Chiapas has fully adopted DOTS (according to official regulations, all patients diagnosed with Pulmonary Tuberculosis (PTB) must be treated under DOTS),⁵ it still presents one of the highest rates of PTB morbidity and mortality in the country. For example, before DOTS was adopted in Chiapas in the year 1996, the PTB mortality rate in 1994 and 1995 was 11.0 and 9.4/100 000, respectively;^6,7 the rate for all the country, was 4.6 and 4.4, respectively. After DOTS, differences still remain between Chiapas and Mexico: during the period 1998–2002, while the PTB mortality rate in Chiapas was 7.1, in the country as a whole was 3.0/100 000 inhabitants.^8–12 In the same period, while the country presented a PTB morbidity rate of 16.6/100 000, in Chiapas it was 34.5,¹³ figures which reveal a deficient implementation of the strategy.

Among the factors making it difficult to adequately implement DOTS in Chiapas, are the lack of health services coverage, low quality of health services, low diagnosis rates for TB, highly rural and dispersed communities, geographical inaccessibility and certain cultural aspects, including language and conceptions about the health-disease process.^14–16

Los Altos Region of Chiapas meets all these conditions. It is one of the poorest regions of Chiapas, with important indigenous settlements and where cultural aspects make it difficult for health workers to communicate sufficiently well with patients to achieve effective follow-up of PTB. Moreover, it is one of the regions most affected by religious, political and armed conflicts.¹⁷ These conflicts have caused continual displacements of the population, to which must be added an escalation in occupation-related migrations, phenomena making control of PTB even more difficult.¹⁷

On the basis of the above considerations, a study was conducted to: analyse the PTB mortality of a cohort of patients in Los Altos Region of Chiapas, who had been diagnosed with PTB between January 1, 1998 (no data are available prior to this date) and December 31, 2002; identify demographic, socioeconomic and health services utilization factors, associated with death from PTB.

Methods

Study area

Chiapas is divided into nine administrative regions. That of Los Altos has a population of 482 010 inhabitants spread among 19 municipalities (Figure 1), 18 of which are considered to be of ‘high-very high’ socioeconomic margination.¹⁸ Only five of its 1226 communities have over 5000 inhabitants. Seventy-one per cent of its population is indigenous (mainly, Mayan ethnic groups), of which one third do not speak Spanish. The illiteracy rate is 42.2%.¹⁸

Figure 1

Situation of Los Altos region within the state of Chiapas, México

Design of the study

Between November 2004 and August 2005, a cross-sectional survey of patients with PTB (PTBP) was conducted. All the patients with a diagnosis of PTB from January 1 to December 31, 2002 (n = 431), were identified from the records of the Health District TB Programme of Los Altos Region of Chiapas. This Health District is in charge of the epidemiological surveillance of TB in the region.⁵

The study considered all PTBP whether diagnosed via AFB smear, culture or clinically, who were aged over 14 years at the moment of diagnosis, independently of their admission category (whether new cases, failures, readmissions or relapses). We reviewed the records of the TB Programme, and we elaborated a patient list, including all information available in the registries: name, sex, age, home address, year of diagnosis and anti-TB treatment outcome.

The study team looked for all patients at their registered home address. Those patients found alive answered a questionnaire, the design of which was based on other instruments previously used in the region.^19,20 For both, alive and deceased patients, we collected information on demographic, socioeconomic and clinical aspects related to PTB, as well as characteristics of their treatment and accessibility to health services. In cases where the patient had died, a verbal autopsy^21,22 was administered to a family member, to obtain additional information about cause, place and date of death, whether or not they had received medical care, whether or not a death certificate had been completed and presence of symptoms compatible with PTB. The verbal autopsy questionnaire used for this study was based on WHO verbal autopsy standards.²³Table 1 shows the results of the search, according discharge categories according to the TB registry of the TB Programme in the studied region.

Table 1

Discharge categories according to the registries of the TB Programme and results of the search for patients in the study

	Discharge category according to the TB Registry of Los Altos region (Chiapas)a
Discharge category n = 431 (%)	Cured 255 (59.2%)	Failed 17 (3.9%)	Defaulted 29 (6.7%)	Died 25 (5.8%)	Transferred 37 (8.6%)	Not specified 68 (15.8%)
Search results
	n = 255 (%)	n = 17 (%)	n = 29 (%)	N = 25 (%)	n = 37 (%)	n = 68 (%)
Alive, n = 208	59.6	29.4	27.6	0	27.0	48.5
Died, n = 78	7.4	35.3	27.6	100.0	29.7	13.2
Not located, n = 145	33.0	35.3	44.8	0	43.3	38.3
Subtotals (%)	100.0	100.0	100.0	100.0	100.0	100.0
Search results in terms of interview response
Alive n = 208
Interviewed, n = 189	140	5	7	0	9	28
Non-response, n = 19	12	0	1	0	1	5
Died n = 78
Verbal autopsy (VA) obtained, n = 55	17	5	8	10	8	7
Located, but no VA (non-response), n = 8	2	1	0	0	3	2
Not located, n = 15b	0	0	0	15	0	0
Situation unknown (Not located), n = 145	84	6	13	0	16	26

	Discharge category according to the PT registry of Los Altos region (Chiapas)c
Discharge category n = 244 (%)	Cured 157 (64.3)	Failed 10 (4.1)	Defaulted 15 (6.1)	Died 10 (4.1)	Transferred 17 (7.0)	Not specified 35 (14.4)
Interviewed n = 189
Alive, n = 184	138	3	7	0	9	27
Alive but ill, n = 5	2	2	0	0	0	1
Verbal autopsy (VA) n = 55
Died due to PTB, n = 40	8	4	8	9	7	4
Died from other causes, n = 15	9	1	0	1	1	3

^aTotal registered patients (n = 431).

^bAccording to the Health District Tuberculosis Registry, these 15 patients figure as deceased, but no cause of death is specified.

^cSearch results by discharge category, among patients located alive and those deceased but with verbal autopsy (n = 244).

Live patients were asked whether they had productive cough. Those who did were asked to provide three sputum samples to be used for smear testing, cultures and if applicable, tests of susceptibility to conventional drug therapy, all such samples being processed according to current Mexican legislation.^5,24 Cultures were processed in the Chiapas State Public Health Laboratories. Drug susceptibility tests were carried out using Bactec MGIT method,²⁵ by the National Epidemiological Diagnosis and Reference Institute. These institutions are officially responsible for conducting the tests involved at state and national levels, respectively.

The statistical analysis consisted of performing tests of association (crude OR, CI 95%), and Analysis of Variance (ANOVA for comparison of continuous variables), all of them were carried out using SPSS version 11.0.²⁶

Ethical aspects

Principles of the Helsinki Declaration²⁷ were followed at all times. Each patient had an informed consent card read aloud to them. Given that according to Mexican health laws²⁸ studies of this type are considered ‘of low risk to health’, approval by a research ethical committee was not required.

Results

The average age of the 431 patients, at the time of their diagnosis, was 39.6 years (SD = 17.4); 54% were males. The patients came from 170 different communities (67% being rural). Of the 295 patients for whom data were available, 254 (86%) were indigenous, 230 (78%) had 0–3 years of schooling, 120 (41%) were engaged in agriculture and 195 (66%) lived in houses with bare earth floors. Table 2 shows the main demographic, socioeconomic, diagnostic and treatment related indicators for the patients studied, contrasting those registered by the TB Programme, with those collected in the home follow-up visits.

Table 2

Comparison of patient characteristics collected from Los Altos Health District Tuberculosis Programme registries, with those collected during the home follow-up visit

Characteristic	Health District TB registries (n = 431)a (%)	Home follow-up visit (n = 295)b (%)
Demographic
Sex
Male	233 (54.1)	155 (52.5)
Female	198 (45.9)	140 (47.5)
Age, complete years
15–34	199 (46.2)	130 (44.1)
35–44	75 (17.4)	54 (18.3)
45 and over	157 (36.4)	111 (37.6)
Indigenous condition
Yes	0 (0.0)	254 (86.1)
No	0 (0.0)	41 (13.9)
Not specified	431 (100.0)	0 (0.0)
Socioeconomic
Educational level (years of schooling)
0–3	0 (0.0)	230 (78.0)
Four and over	0 (0.0)	65 (22.0)
Not specified	431 (100.0)	0 (0.0)
Type of occupation
Agriculture	0 (0.0)	120 (40.7)
Non-agriculture	0 (0.0)	175 (59.3)
Not specified	431 (100.0)	0 (0.0)
House flooring material
Bare earth	0 (0.0)	195 (66.1)
Some form of covering	0 (0.0)	100 (33.9)
Not specified	431 (100.0)	0 (0.0)
Fuel used for cooking in the home
Wood or coal	0 (0.0)	259 (87.8)
Gas	0 (0.0)	36 (12.2)
Not specified	431 (100.0)	0 (0.0)
Indicators from the TB Programme
Admission category
New case	395 (91.6)	–
Relapse	15 (3.5)	–
Readmission	17 (3.9)	–
Failure	4 (0.8)	–
Correct home address allowed location of patient
Yes	268 (62.2)	–
No	163 (39.8)	–
Anti-tuberculosis treatment outcome
Cured	255 (59.2)	–
Failed	17 (3.9)	–
Defaulted	29 (6.7)	–
Died	25 (5.8)	–
Referred	37 (8.6)	–
Not specified	68 (15.8)	–
Outcome according to home follow-up visit
Patients alive	–	208 (48.3)
Patients dead	–	78 (18.1)
Not located	–	145 (33.6)
Status of PTB (% only patients alive)
Cured	–	184 (88.5)
PTB-positive	–	5 (2.4)
Unknown (non-response)	–	19 (9.1)

^aRegistries of the Los Altos Health District Tuberculosis Programme, 1998–2002.

^bResults of the search to locate the 431 subjects of the study in their homes, during the period November 2004 to August 2005. It includes information from: 189 patients interviewed, 55 verbal autopsies provided by family members of deceased patients and 51 patients not located but for whom socioeconomic and demographic characteristics were available from the research project: Sánchez-Pérez HJ, Martín-Mateo M, Jansá JM. ‘Tuberculosis Pulmonar en Los Altos de Chiapas: Avances o Retrocesos en su Control’. San Cristóbal de Las Casas, Chiapas: El Colegio de la Frontera Sur, 2001.

With regard to the records of the PTB Programme in the region, these were incomplete and erroneous in many cases. For example, 16% had not specified the discharge category, 40% had not the right home address, none had information about whether they were indigenous or not, among others characteristics (Tables 1 and 2).

The results of home follow-up visits were categorized as follows:

1. Alive and interviewed (n = 189): Those patients found alive, and who agreed to participate.

2. Died from PTB (n = 40): Deceased patients for whom the cause of death was associated with PTB. This included: 15 cases in which the family, during the verbal autopsy, attributed the death from PTB, and these cases were subsequently corroborated by the death certificate (i.e. the death certificate explicitly mentioned PTB—five of them were not registered by the TB Programme), and 25 cases where the family attributed the death from PTB, 21 of which had no death certificate, while the death certificates of four gave some other cause, but there were evident classification errors (two gave the cause as ‘cardio respiratory arrest’, one ‘multiple organ failure’ and one ‘respiratory insufficiency and severe malnutrition’). In these four cases, according to information provided by the family, respiratory symptoms compatible with PTB had been observed. The level of agreement, measured by Kappa Index,²⁹ between the cause of death confirmed by death certificate, and that attributed by the family in the verbal autopsy, was 0.61 (Table 3).

3. Died from cause other than PTB (n = 15): Patients whose cause of death in the death certificate was different from PTB, or whose family indicated that the patient did not present respiratory symptoms at the moment of death.

4. Cause of death unknown (n = 23): Eight patients who had died according to the TB Programme registry, although cause was not specified, but whose families refused to participate (family members said: ‘what is the point, since he is dead’), and 15 patients who could not be located (nor their families).

5. Non-response (n = 19): Patients who were located alive, but refused to participate. The main reasons given were: ‘I am well now, so no need to participate’.

6. Not located (n = 145): Patients not found— 45 due to having migrated, 92 because of lack of a correct home address. In this group, there were 15 out of the 25 in the discharged ‘Died’ category by the TB Programme.

Table 3

Cause of death according to the death certificate compared to cause of death according to verbal autopsy^a

	Symptoms compatible with PTB at the moment of death and death attributed to PTB by the family (verbal autopsy)
	Yes	No	Total
PTB	15	0	15
Other than PTB	4b	5	9
No death certificate	21	10	31
Total	40	15	55

^aKappa Index was 0.61 for concordance between ‘Cause according to death certificate’ and ‘Symptoms compatible with PTB at the time of death and death attributed to PTB by the family (Verbal Autopsy)’.

^bCauses of death cited in death certificates were: cardio-respiratory arrest (2), multiple organ failure (1), respiratory insufficiency and severe malnutrition (1).

Thus, of 431 initial cases, for the purposes of the study, two of the groups described above were used (n = 229): 40 patients who died from PTB, and 189 located alive and interviewed.

Of these 189 patients, 110 (58%) no longer had any respiratory symptoms, while 79 (42%) had such symptoms, 52 of them (66%) provided one or more sputum samples. Of these 52 patients, the TB Programme considered 36 (69%) as cured, 4 (8%) as failures, 7 (14%) defaulted from treatment and 5 (10%) had no treatment outcome recorded. Of those 27 who did not provide sputum, 17 (63%) figured in the Programme as cured and 10 (37%) had no outcome recorded.

Of the 52 patients who provided sputum samples, five were PTB-positive (three were multi-drug resistant, with histories of at least two prior treatments, but none under DOTS). The categories of admissions and discharges to/from the TB Programme of the three multi-drug resistant cases were: one entered as a readmission and was discharged as cured; one entered as a readmission and was discharged as a failure; the third was classed as a failure both at admission and at discharge.

Vital status (alive vs died due to PTB) according to indicators analysed

Demographic and socioeconomic characteristics of the patients

No differences were found in terms of sex, or being indigenous, or monolingual, between patients who died from PTB (‘PTB-deaths’) and those who were located alive. The age was the only characteristic of this group associated with death from PTB (Table 4).

Table 4

Socioeconomic, demographic, and anti TB-treatment indicators, associated with the fact of having died from PTB^a

Indicator	Died from PTB (%)	Alive (%)	OR (95%CI)
Demographic
Age
45 and over (n = 75)	18 (24.0)	57 (76.0)	1.9 (0.9–3.8)
15–44 years (n = 154)	22 (14.3)	132 (85.7)
Socio economic
Educational level (years of schooling)
0–3 (n = 175)	36 (20.6)	139 (79.4)	3.3 (1.1–9.5)
Four and over (n = 54)	4 (7.4)	50 (92.6)
Type of occupation
Peasant (n = 95)	23 (24.2)	72 (75.8)	2.2 (1.1–4.4)
Other (n = 134)	17 (12.7)	117 (87.3)
Residing in capital village of their municipality
No (n = 154)	33 (21.4)	121 (78.6)	1.1 (1.0–1.3)
Yes (n = 75)	7 (9.3)	68 (90.7)
Place of residence
Rural (n = 162)	34 (21.0)	128 (79.0)	2.7 (1.5–6.8)
Urban (n = 67)	6 (9.0)	61 (91.0)
Anti-tb treatment
Treated under DOTS
No (n = 194)	38 (19.6)	156 (80.4)	1.2 (1.1–1.3)
Yes (n = 35)	2 (5.7)	33 (94.3)
Defaulted (remained in treatment <6 months)
Yes (n = 56)	27 (48.2)	29 (51.8)	11.5 (5.3–24.8)
No (n = 173)	13 (7.5)	160 (92.5)
Readmitted after having defaulted
Yes (n = 12)	0 (0.0)	12 (100.0)	2.6 (1.8–3.8)
No (n = 56)	34 (61.4)	22 (38.6)

^aVariables analysed without association with the response variable: Socio-demographic: Sex: male (n = 114) vs Female (n = 115): 21.1 and 13.9% of deaths, respectively; with social security: yes (n = 6) vs no (n = 223): 0.0 and 17.9% of deaths; Overcrowding yes (n = 107) vs no (n = 122): 15.9 and 18.9% of deaths; cooking gas facilities: yes (n = 20) vs no (n = 209): 5.0 and 18.7% of deaths; with floor covered: yes (n = 73) vs no (n = 156): 12.3 and 19.9% of deaths; with electricity: yes (n = 219) vs no (n = 10): 17.4 and 20.0% of deaths. Access to health services: Distance from home to the nearest health facility: <30 min (n = 144) vs 30 min and over (n = 85): 18.8 and 15.3% of deaths. With health facilities in the community of residence: yes (n = 118) vs no (n = 111): 0.0 and 17.9% of deaths; Clinics: Concomitants diseases at the moment of diagnostic of PTB: yes (n = 24) vs no (n = 205): 25.0 and 16.6% of deaths; Familiar antecedents of PTB: yes (n = 6) vs no (n = 223): 0.0 and 17.9% of deaths; Anti-TB treatment: With previous antecedent of anti-TB treatment: yes (n = 31) vs no (n = 198): 22.6 and 16.7% of deaths; Health care provider who made the diagnosis of PTB: government (n = 213) vs no government (n = 16): 17.8 and 12.5% of deaths; denial of health care services: yes (n = 20) vs no (n = 209): 25.0 and 16.7% of deaths.

Regarding socioeconomic indicators, no differences were found for housing conditions, social security or distance of the community from the principal town of the region (i.e. San Cristobal de Las Casas). The indicators associated with the response variable were: years of schooling, type of occupation, residing in the capital village of the municipality and place of residence (Table 4).

Anti-TB treatment and other health services utilization indicators

Between patients who died due to PTB (PTB-deaths) and patients located alive, there were no differences with respect to either year or agent of diagnosis, whether or not they were a new PTB case at the time of admission to the TB Programme, whether there was any history of TB in their family, whether diagnosed with concomitant diseases (at the time of PTB diagnosis), denied consultations by the health services, presence of a medical facility in their community of residence, nor the distance to the nearest medical unit. Conversely, differences were found for the following indicators (Table 4): not having been treated under DOTS, defaulting from treatment (the main causes of defaulting treatment appear in Table 5), and not having been readmitted to anti-TB treatment after having defaulted.

Table 5

Main reasons given for defaulting from anti-TB treatment

Main reasons given for defaulting from anti-TB treatment	Died from PTB (n = 21)a (%)	Alive (n = 29) (%)
‘Felt better, didn't want to go to the clinic’	9 (42.9)	16 (55.2)
‘Felt worse with the treatment’	2 (9.5)	5 (17.2)
‘Shortage of medicine at the clinic’	2 (9.5)	3 (10.3)
‘Alcoholism’	4 (19.0)	1 (3.4)
‘Medical advice’	1 (4.8)	2 (6.9)
Other reasonsb	3 (14.3)	2 (6.9)

^aSix patients who died before completing their anti-TB treatment were not included in this category.

^bIncludes five cases: three died, two of whom felt that ‘the pills were not doing anything’, and one who preferred to take the herbs prescribed by a traditional-medicine doctor; two still alive: one preferred to take herbs prescribed by traditional-medicine doctor, one could not afford to go to the health services clinic.

Indicators of the circumstances of the death among deceased patients

Table 6 presents the main indicators of the circumstances of death of 40 PTB-deaths, along with those of another 15 who died from other causes.

Table 6

Indicators of the circumstances of death

	Died from PTB (n = 40)	Died from other causes (n = 15)	OR (95% CI)
Age at death
Average (SD)	44.95 (15.54)	51.6 (17.8)
Aged 15–44 years (%)	22 (55.0)	5 (33.3)	2.4 (0.7–8.4)
Aged 45 and overa (%)	18 (45.0)	10 (66.7)	–
Percentage of time elapsed between PTB diagnosis and death
<1 year	15 (37.5)	0 (0.0)	1.5 (1.0–2.5)
From 1 year to <2 yearsa	7 (17.5)	4 (26.7)	–
Two years or more	18 (45.0)	11 (77.3)	1.1 (0.3–4.5)
Percentage of place of death
At home	34 (85.0)	8 (53.3)	2.9 (2.2–6.9)
In a public place (Street, road, on the field)	2 (5.0)	2 (13.4)	0.8 (0.8–8.5)
In health facilitiesa	4 (10.0)	5 (33.3)	–
Percentage of with medical care at the time of death
No	33 (82.5)	10 (66.7)	1.8 (0.8–4.2)
Yesa	7 (17.5)	5 (33.3)	–
Percentage of with death certificate
No	7 (17.5)	0 (0.0)	0.7 (0.6–0.8)
Yesa	33 (82.5)	15 (100.0)	–
Able to show death certificate (of the who claimed to have death certificate) (%)
No	15 (55.0)	10 (66.7)	2.4 (0.7–8.6)
Yes	18 (45.0)	5 (33.3)

^aReference group.

Discussion

The demographic and socioeconomic indicators of the study population reveal that it consists largely of indigenous people living in conditions of marked poverty. Among the main findings, it should be pointed that: the records of the PTB Programme in the area were incomplete and erroneous in many cases; 78 (18%) of those searched at their homes had died, 40 of the deaths apparently were associated with PTB. Of these 40, 33 died without having received any medical care. The factors associated with dying from PTB were: being 45 years and over, having 0–3 years of schooling, being a peasant, not living in the capital village of their municipality, living in a rural community, not having been treated in DOTS and having defaulted from treatment.

The analysis of TB Programme registries showed that: patient information in the records is deficient; many cases lacked address (38% had no address or other references to locate the patient), or treatment outcome (16% had no treatment outcome, a further 9% figured as ‘transferred’ generally meaning the TB Programme had ceased to have any contact with them). Similarly, given that the region is inhabited by both non-indigenous and several indigenous ethnic groups, and that an important proportion live in extreme poverty, it would be advisable for the registers to contain information about ethnicity, socioeconomic aspects (schooling, occupation, etc.) and health (concomitant health conditions, etc) of the patients. The fact that the registers lack address details, is in itself an indication that DOTS is not being carried out satisfactorily in the region studied. While this lack of information is not solely attributable to the health services, but also to the patients themselves (e.g. providing an incorrect address), the Programme ought to have mechanisms to verify these aspects. Furthermore, the cure rate found (of 59%) falls short from the 85% recommended by WHO.³⁰

With regard to the household interview results, we would like to stress the following findings: the high number of patients not located (34%), and the high mortality among the patients who were found. Given that among the studied patients we found that a higher mortality rate was associated with low schooling, being a peasant, living in a rural community and not living in the capital village, and that the study region is mainly rural and agricultural, with the highest level of illiteracy in Chiapas, it is very likely that among the group of patients not found, the mortality rate were as high as that of the located patients. In addition, the low accessibility to the health services in this region, including the poor access to the DOTS strategy, contributes to the plausibility of this hypothesis.

It is noteworthy that 55% of those patients, whose death was related to PTB, had died within two years of being diagnosed. Two aspects may be responsible for this situation: the possible delays with which diagnoses and anti-TB treatment are carried out in the region; and the possible inefficiency of the health services. Chiapas is the Mexican state with the highest PTB mortality.¹² Whereas the life expectancy in Chiapas is 72.2 years,¹⁸ the average age of deceased patients was 47.4 years, representing an average of almost 25 potential years of life lost, along with the family and socioeconomic repercussions which this implies. Besides, the majority of these patients died without receiving any medical care: in 23 cases of the 78 who died, lack of data meant that neither the cause nor circumstances of death could be determined. However, in patients for whom such information was collected, those who died from PTB appear to have a death in worse conditions (they died younger, at their homes, without any medical care and sooner once they have been diagnosed) than people dying from other causes (Table 6), something which reveals a necessity to conduct studies in this type of patient analyzing the role played by lack of access to health services, as well as cultural aspects, such as beliefs and preferences in the use of traditional medicine services, as other authors have suggested.^14–16,31

It should be noted that of the 55 documented deaths, 72% were very probably due to PTB, considerably more than the 42% mortality found by a study performed in another Mexican state.³² Unfortunately, it was only possible to corroborate the cause of death via death certificate in 15 cases. Nevertheless, two aspects are to be noted: the ‘acceptable’ degree of concordance between the probable cause of death according to the family and that found in the death certificates (Kappa Index of 0.61),²⁹ when this was possible; and, although the cause may have been expressed in the death certificate, this is not necessarily a guarantee of the quality of diagnosis, since in rural areas of Chiapas (and particularly in the region studied) the infrastructure necessary to confirm diagnoses made by health personnel is lacking: Chiapas has the highest proportion of misclassified deaths in the country.¹²

On the other hand, it is possible that the high mortality found could have had an influence in the finding of a PTB positivity rate under 3%. Note however that, of the 79 productive coughers found, it was only possible to obtain sputum samples from 66% (and most of them of poor quality), something which could have limited the capacity to identify more cases of active PTB. The low number and quality of sputum samples could possibly be due, as noted by other research done in areas with similar conditions,³³ to communication barriers between researchers and patients, arising from language and cultural differences, as well as from the low educational level of the population. Even so, it is important to note that of the five PTB cases found, three presented multi-drug resistance.

With regard to demographic indicators associated with death from PTB, it is noteworthy that a higher age was associated with higher PTB mortality, tending to suggest that among older patients, the accumulation of unfavourable living conditions (i.e. malnutrition and poverty) together with the probably deficient medical care by the health services, make them an especially vulnerable group. Chiapas is the state with the lowest life expectancy in the country, and the highest level of socioeconomic margination.¹⁷ With respect to whether the indigenous population was more affected by PTB than the non-indigenous population, given that the results obtained are not population-based, and that the registries of the Ministry of Health gather no information on this aspect, further studies are required to analyse this issue. Of the socioeconomic indicators analysed, it is particularly notable that those of a ‘socio-environmental’ nature (i.e. housing conditions) were not associated with higher mortality, while variables depending more on the individual, such as educational level, and occupation, were associated. This could be due to the fact that the socio-environmental conditions are relatively homogeneous among the patients studied, whereas the more individual-related variables are less so. Patients with higher educational level, residing in an urban setting, and not involved in agriculture, could be in a better position, both economically and in terms of information and in consequence, be more capable of taking decisions and of obtaining more and better care, than their counterparts.

With respect to the proportion of patients who received anti-TB treatment under DOTS, this was very low. Given that treating PTB patients according to DOTS is essential to controlling the disease,^32,34,35 omitting to do so is a very serious failing of the health services deserving to be addressed immediately.

Finally, among the main limitations of this study, it should be pointed out that certain aspects could not be analysed due to absence of data corresponding to patients who died: their drug multi-resistance condition, more details about their treatment and whether or not the armed conflict in Chiapas,¹⁷ affected their treatment. In addition, there may be a possible selection bias, since 34% of the patients, were not found.

Conclusions

The high PTB mortality rate found constitutes a serious public health problem in the region, deserving special attention. The results show that there are serious deficiencies in its TB Programme. Within this context, those aged 45 years and over, with lower educational levels, engaged in agriculture and residing in rural communities, present a higher chance of dying after diagnosis of PTB. It is necessary to evaluate and restructure the TB Programme in the study region, and in other areas with similar conditions, not just the information collected/recorded, but all aspects related with application of DOTS.

Acknowledgements

The authors are grateful to Julio Cesar Arias and Adalberto Díaz, for their valuable contributions in the field work, and to Dave McFarlane for his invaluable contributions. Thanks also to the Chiapas State Science and Technology Council for the financial support, and to the Chiapas Ministry of Health, and the InDRE, for their help and support in conducting this study.

Conflict of interest: None declared.

KEY MESSAGES

• The results show that there are serious deficiencies in Health District TB Programme.

• Seventy-eight (18%) of those searched at their homes had died, 40 of the deaths apparently were associated with PTB. Of these 40, 33 died without having received any medical care.

• The high PTB mortality rate found constitutes a serious public health problem in the region, deserving special attention.

• Being 45 years and over, having 0–3 years of schooling, being a peasant, not living in the capital village of their municipality, living in a rural community, not having been treated in DOTS and having defaulted from treatment, present a higher chance of dying after diagnosis of PTB.