A qualitative study of factors resulting in care delays for adults with meningitis in Zambia

Abstract

Background

Meningitis causes significant mortality in regions with high comorbid HIV and TB. Improved outcomes are hindered by limited understanding of factors that delay adequate care.

Methods

In-depth interviews of patients admitted to the University Teaching Hospital with suspected meningitis, their caregivers, doctors and nurses were conducted. Patient/caregiver interviews explored meningitis understanding, treatment prior to admission and experiences since admission. Provider interviews addressed current and prior experiences with meningitis patients and hospital barriers to care. A conceptual framework based on the Three Delays Model identified factors that delayed care.

Results

Twenty-six patient/caregiver, eight doctor and eight nurse interviews occurred. Four delays were identified: in-home care; transportation to a health facility; clinic/first-level hospital care; and third-level hospital. Overcrowding and costly diagnostic testing delayed outpatient care; 23% of patients began with treatment inside the home due to prior negative experiences with biomedical care. Admission occurred after multiple clinic visits, where subsequent delays occurred during testing and treatment.

Conclusions

Delays in care from home to hospital impair quality meningitis care in Zambia. Interventions to improve outcomes must address patient, community and health systems factors. Patient/caregiver education regarding signs of meningitis and indications for care-seeking are warranted to reduce treatment delays.

Introduction

Meningitis remains a significant cause of mortality in sub-Saharan Africa.^1–3 Inpatient mortality can be up to 70% depending on etiology.^4–8 Among survivors, residual deficits are the second greatest contributor to neurologic disability-adjusted life years for Eastern sub-Saharan Africa.^9,10 For these reasons, the WHO launched an initiative aimed at decreasing the mortality and morbidity of meningitis by 2030.¹¹ However, progress depends on identification of reasons for poor outcomes, including barriers to adequate meningitis care.

Both symptom severity and late care presentation are associated with mortality.^12,13 For patients with bacterial meningitis, qualitative studies shed light on treatment-seeking behavior. Malawian families sought care when the individual could not eat or work,¹⁴ whereas in Nigeria household income influenced the timing and type of care sought.¹⁵ However, for more insidious meningitis causes, such as tuberculous or cryptococcal meningitis, factors influencing care-seeking behaviors have not been elucidated. Further, while delayed antibiotic administration is associated with increased mortality in bacterial meningitis,¹⁶ little is known regarding other health systems barriers that could delay care and influence outcomes.

Zambia is a lower-middle income country of approximately 18 million people with an adult HIV prevalence of 11.3% and TB prevalence of nearly 1%.^17,18 Delays in TB diagnosis and treatment are common, and up to 10% of cases may include meningitis.^19,20 During the 2010–2012 nationally representative verbal autopsy study, meningitis caused 1.6% of adult deaths.²¹ Meningitis mortality in Zambia is higher than expected for its sociodemographic index, which suggests additional factors influence outcomes.⁹

To identify factors that delay adequate care for patients with suspected meningitis, we examined the treatment pathways for 26 patients admitted to the University of Zambia University Teaching Hospital (UTH)—Adult Hospital in Lusaka. We report patterns in outpatient care-seeking and examine diagnostic testing and inpatient treatment from the perspective of patients, caregivers, doctors and nurses.

Materials and Methods

Research setting

This study was conducted from January to August 2020. The UTH is a 1600-bed third-level hospital that serves as Zambia's national referral hospital. As part of the national subsidized healthcare system, patients do not pay for daily bed use, basic laboratory testing or medications in the hospital pharmacy. However, advanced investigations, including neuroimaging, were not covered at the time of this study.

While UTH patients are from across Zambia, most reside in Lusaka District, a city of 2.7 million people.²² Lusaka's adult HIV prevalence is >15%.¹⁷ The socioeconomic status and housing conditions of residents are mixed, with the majority working in the informal economic sector or low-paying jobs in the public or private sectors.¹⁷

Health services are mainly provided by public sector facilities. Clinic catchment areas range from 60 000 to >100 000 people and provide basic ambulatory and preventative services.²³ Facility overcrowding is common.^24,25 First-level district hospitals provide ambulatory care but can admit patients, if needed.²³ Second- and third-level hospitals, like UTH, provide more specialized services.²³ Private facilities are available and refer patients to public health facilities including UTH. There are traditional healers as well as religious organizations, predominantly Christian, which play a role in residents’ health-seeking behaviors.

Research design and study participants

An in-depth qualitative study was conducted with patients, caregivers, doctors and nurses at the UTH-Adult Hospital. Purposive sampling was used to select participants. Interviewers also completed field notes describing observations and experiences conducting interviews.

Patients and caregivers

Eligible patients were aged >18 y admitted with suspected meningitis. A physician with additional training in neurology (MAE, MC, AMM) went bed-to-bed in the emergency, admission and medicine wards to identify patients. Study staff were not involved in the care of eligible patients. Once an eligible patient was identified, they were approached for informed consent. If the patient was unable to consent, a relative who resided with the patient was identified for participation. Once written consent was obtained, participants were interviewed at the patient's bedside or in a private room. After 18 March 2020, all interviews were conducted via telephone to promote social distancing during the coronavirus disease 2019 (COVID-19) pandemic. At the end of the interview, participants were provided with 50 Kwacha (∼US$\mathrm{\$}$4) for transportation costs.

Interviews were conducted in the language of the participant's choice by a trained interviewer (MM, GM, NN) using a guide. Participants were asked to provide a detailed illness narrative prior to presenting to UTH. They were also asked about their experiences since admission, including testing performed and treatment initiated. To provide context, participants described their area of residence, including healthcare access and prior illness experiences.

Doctors and nurses

Once a patient/caregiver interview was completed, an eligible healthcare provider was identified. Eligible providers were doctors or nurses providing care to a consenting patient/caregiver. Interviews were conducted in a private room or via telephone. All were provided with 50 Kwacha (∼US$\mathrm{\$}$4).

Providers were asked to describe their daily hospital responsibilities. Interviews then focused on participants’ experiences with meningitis patients early in their training before transitioning to the current patient. Participants were asked to detail their approach to educating patients about their illness and challenges inherent to caring for meningitis patients.

Analysis and conceptual framework

Interviews were audio recorded, with participant permission, then translated and transcribed into English. Interviews were coded iteratively with ATLAS.ti v7 (ATLAS.ti Scientific Software Development GmbH, Berlin, Germany) to identify factors influencing healthcare behavior in the outpatient and inpatient settings.²⁶ Interviews were coded both by the interviewer (MM, GM, NN) and a senior qualitative researcher (MAE, BC) and disagreements were resolved by consensus. Field notes corroborated identified themes.

A conceptual framework based on the Three Delays Model categorized themes. The Three Delays Model, originally applied to maternal mortality, proposes that poor patient outcomes are associated with delays in: deciding to seek care; reaching the appropriate facility; and receiving adequate care.²⁷

Patient consent

Written, informed consent was obtained in English or a local language.

Results

The characteristics of the study participants are provided in Table 1. Twenty-six patient case studies were included, encompassing interviews with 25 caregivers, 1 patient, 8 physicians and 8 nurses. Eight caregivers (31%), seven (89%) doctors and five (63%) nurses were interviewed via telephone. Interviews ranged in duration from 15 to 90 min. Physician interviews often occurred between hospital shifts and, due to competing clinical demands, were often shorter than patient/caregiver and nurse interviews. All patients/caregivers consented to audio recording whereas three providers did not.

Caregivers were the patient's spouse (28%), mother (27%) or sister (19%). One-third (9/26; 35%) had completed up to grade seven of education, 46% were unemployed and 65% of patients were people living with HIV (PLWH). While eight (31%) patients had been ill for <1 wk, six (23%) had been ill for >2 mo. Although final diagnoses were not collected for all patients, 42% (11/26) were suspected to have tuberculous meningitis based upon record review at the time of the interview.

Patient illness episodes averaged three treatment steps prior to admission (range 1–7). Treatment steps were classified as home treatment with modern pharmaceuticals, informal community services (with a traditional healer, herbalist or prayer) or formal biomedical services at a public or private facility. In 23% (6/26) of illnesses, initial treatment occurred inside the home. Repeat clinic visits were common; only 31% (8/26) were referred to a higher level of care the first time they were evaluated.

The following four stages of delay were identified by participants: (1) in the home/decision to seek care outside the home; (2) transportation to health facility; (3) first-level health facility; and (4) third-level referral health facility.

Delay #1: In the home/decision to seek care outside of the home

Recognition of meningitis symptoms: Ten participants began treatment inside the home. Perceived symptom severity had the greatest impact on choice of treatment. Seven of 26 (27%) patients’ first symptom was headache. Headache was attributed to dehydration, high blood pressure or carrying cold objects, such as popsicles, on top of the head. As a result, patients initiated over-the-counter pain medications.

Prior illness experiences affected interpretation of symptoms. Three participants reported painful legs, which they associated with possible TB infection even in the absence of cough. When headache was accompanied by fever or chills, symptoms were attributed to malaria. Most noted that malaria was increasingly uncommon within Lusaka but was suspected if one had traveled. Despite this, participants purchased over-the-counter antimalarial medications even in the absence of recent travel. Meningitis and severe malaria were considered synonymous:

I just know that it is malaria that is out of control. Malaria that has changed its level and it is untreatable and then turns to meningitis (32-y-old female, spouse to a male patient).

Alarming symptoms, particularly those inconsistent with malaria, or a lack of improvement despite over-the-counter pharmaceuticals, prompted care-seeking outside of the home. Confusion, convulsions and hallucinations were universally regarded as severe symptoms:

For days, he was just complaining of headache. Once he takes a pain killer. Then he just stopped talking. You would be talking to him, but he was not responding. That's when we said, ‘let's take him to the hospital’ (63-y-old female, mother to a male patient).

Avoidance of government clinics: All participants indicated that a government or private health facility was the most appropriate choice when a patient exhibited meningitis symptoms. However, they were hesitant to go to a government clinic due to overcrowding and most lacked the funds for a private clinic. All participants reported prior experiences using over-the-counter medications or home remedies for treatments. Three participants sought traditional health services during this illness episode. All continued over-the-counter medications while using prayer and herbal medicines.

Delay #2: Transportation to health facility

Most patients/caregivers walked to the nearest biomedical facility for care. These were frequently government clinics, although one caregiver attended a private clinic. When the patient was too ill to walk, participants booked a taxi or sought assistance from a relative with a vehicle. Transportation costs, which ranged from ZMW 30 to 200 (US$\mathrm{\$}$2–16), and availability contributed to delays. Caregivers delayed transferring the patient to a health facility until a relative was available to accompany and stay with the patient. Because of insufficient staffing at clinics and hospitals, relatives are often required to provide basic care needs such as feeding, hygiene and medication administration.

Delay #3: clinic and first-level hospital

Prolonged wait for care: After arrival at the health facility, participants reported a prolonged wait until they could be seen:

You could be there at 08 hours and maybe you will see the doctor at 16 hours, just there waiting. The next time you get sick you just say, ‘no let me stay away’ (55-y-old female, mother to a female patient).

One caregiver sought care at her daughter's HIV clinic instead of the general outpatient clinic. There, staff recognized the patient, and treatment was expedited.

Clinic providers gave medications to 69% (18/26) of patients on their first visit. Many received over-the-counter painkillers for headache without further diagnostic testing. Participants viewed this unfavorably as they received little additional benefit from seeking care at the clinic.

A clinic is like a first point of call, and they will dilly-dally, dilly-dally. Just give you a few pills, you go back, you think everything is okay (26-y-old male patient).

Testing as a barrier to care: In some cases, malaria testing or a chest x-ray were required prior to treatment. However, this was atypical. Two caregivers could not afford testing and, as a result, the patients did not receive treatment despite being seen by a provider. Two caregivers reported that testing was a barrier to care as their male patients believed it would include unwanted HIV screening. Participants viewed testing prior to treatment positively as they believed it would lead to a diagnosis. However, delays in receiving test results were common. One caregiver reported that a lumbar puncture (LP) for cerebrospinal fluid testing was performed at a first-level hospital with the understanding that the results would be sent to UTH; however, the results were subsequently lost.

Repeated clinic visits required before transfer: Thirteen of 26 patients (50%) required >2 clinic visits before referral for inpatient evaluation. Participants were generally advised to follow up in 1 wk but returned sooner due to worsening symptoms. One participant did not return to the clinic and instead brought his wife directly to UTH. However, two female caregivers returned to the clinic first as they anticipated the hospital would not evaluate their patient without a clinic referral. This may be rooted in the fact that second- and third-level hospitals may charge user fees to patients who bypass primary clinics.²³

Delay #4: third-level referral hospital

Testing as a barrier to care: After arriving at UTH, the first care delay occurred during diagnostic testing. LP and neuroimaging were commonly ordered. Caregivers and providers agreed that LP was often refused by patients/caregivers out of concern for a heightened risk of death during the procedure:

What is in the minds of people is that when you do the lumbar puncture, it's 50/50. Either the patient dies or luckily survives (42-y-old male, spouse to a female patient).

Caregivers often agreed to neuroimaging instead of LP. Most requested assistance from the hospital's social welfare office for neuroimaging costs that resulted in further delays. When the social welfare office did not help, caregivers sought assistance from relatives or declined testing. Among patients that completed testing, caregivers and providers reported delays in receiving results or receiving incomplete results that adversely impacted care:

Depending on who was working in the lab the previous day, you might not get your results. Then some other investigations the lab can't run them, so you are forced to ask the patient if they can carry out the investigations from outside the hospital. The majority of people can't even afford a meal a day (31-y-old female physician).

Doctors generally expressed confidence in laboratory results. However, one physician indicated she would consider repeating an LP if she was concerned the results were incorrect.

Selecting appropriate medications: The second inpatient care delay occurred when selecting appropriate medications. Doctors initiated antibiotics immediately for patients with suspected meningitis. However, when LP was refused, they relied on signs of clinical improvement to guide treatment. If an LP was performed after initiating antibiotics and did not reveal a pathogen, one doctor continued antibiotics for 14 d whereas others halted, or modified treatment based on results. All doctors and nurses recalled medication stockouts that threatened appropriate treatment of patients, although this was not a barrier at the time of the interviews.

Impact of patient/caregiver understanding: Patient/caregiver understanding of disease impacted all aspects of care, including diagnosis and treatment. All participants agreed that the community's understanding of meningitis was limited. Doctors and nurses considered patients/caregivers who defined meningitis as an infection of or an attack on the brain to have adequate understanding. However, only 15% (4/26) of patients/caregivers associated meningitis with the brain or an infection besides malaria. Ten out of 26 (38%) did not know what meningitis was, although some associated it with HIV infection. Two defined meningitis as neck stiffness (chinyamankazi). In all cases, suspicion for meningitis was not raised until the patient sought care in the formal health sector. Only 15% (4/26) of patients/caregivers were told about meningitis prior to admission.

Providers thought it was important to explain illnesses to patients/caregivers; however, they noted multiple barriers. They cited insufficient time due to competing clinical demands as well as difficulty in identifying interpreters to educate patients who did not speak Bemba, Nyanja or English. Caregivers’ lack of understanding caused uncertainty during the patient's hospitalization:

Nobody told us what to expect. Even right now we don't know how long he will be in the hospital. We just have to guess (26-y-old female, spouse to a male patient).

Poor caregiver understanding impacted treatment. One caregiver considered discharge against medical advice to seek care from a traditional healer. Three of 26 caregivers (12%) reported that their patient was not being treated. Upon further investigation, all were receiving medications; however, caregivers denied a treatment plan being explained to them.

Discussion

While late presentation for biomedical care has been reported among patients with meningitis^14,28,29 and is associated with increased mortality,¹² this is a culmination of several delays in care. In our setting, patients with meningitis must reach the highest level of biomedical care before diagnostic investigations and treatments are initiated and up to four delays occur prior to admission. Accurately characterizing and addressing care delays is essential to improving outcomes.

We found that initial meningitis symptoms were managed symptomatically with lower-cost over-the counter treatments available in the home. While biomedical clinics were considered the best care for patients, dissatisfaction with government clinics due to overcrowding, perceived incompetence of providers and high testing costs was common, corroborating prior findings in this setting.³⁰ Reticence to obtain services from these clinics delayed care. Medical pluralism, when patients simultaneously seek care from formal biomedical and informal providers, has been reported by patients with bacterial meningitis in Burkina Faso and Nigeria.^15,31 By recruiting participants from UTH, we were unable to corroborate patient experiences prior to admission and may have inadvertently underestimated the number of patients seeking informal care.

Diagnosing meningitis requires that providers recognize clinical signs and symptoms of meningitis and obtain the appropriate diagnostic testing. While we did not interview providers at local clinics and first-level hospitals, earlier symptom recognition at that level may expedite care for meningitis patients. Further, caregivers in our study referred to UTH complied with recommendations for transfer from the lower-level health facility to the referral hospital, likely due to the severity of the patient's symptoms, and ambulance transfer occurred on the day of request. However, many more patients may forgo this recommendation or remain at the first-level hospitals for care, and these patients would not have been captured in our study. Given that appropriate diagnostic testing often did not occur until a patient reached UTH, this may further exacerbate poor outcomes. Examining factors that contribute to care delays in clinics and first-level hospitals is warranted so interventions can appropriately target delays prior to admission.

While approximately half of our patients were treated empirically for tuberculous meningitis, this diagnosis was not verified due to difficulties in obtaining appropriate testing. One-third of South African PLWH undergoing LP for suspected meningitis were diagnosed with TB.³² Tuberculous meningitis is underdiagnosed in settings like Zambia, and diagnostic challenges contribute to high mortality rates,³³ which are further compounded by delays in treatment.^19,34 In addition to implementing community-based TB case-finding strategies, interventions to improve patient and caregiver recognition of meningitis symptoms, particularly among PLWH seeking routine clinical care, are warranted to reduce delays.

Interviewed providers were employed at the highest level of care and, while they did not endorse difficulty in identifying patients with possible meningitis, they noted multiple barriers to diagnostic testing. LP is an essential neurodiagnostic procedure in patients with meningitis and, while low LP uptake has been reported,³⁵ our findings highlight the challenges posed by low LP uptake. Antibiotics may sterilize the cerebrospinal fluid, decreasing the likelihood of pathogen identification when an LP is delayed.^36,37 In our setting, where caregivers reconsider LP after initial refusal,²⁹ doctors reported divergent clinical practices that may adversely impact outcomes when making treatment decisions using results obtained after antibiotic administration. Interviewed providers were often the patient's primary doctor as identified by the medical record and were often junior physicians early in their training. This may have inadvertently limited the breadth of knowledge and experience shared by providers. While consultants are not often the first point of contact for patient care decisions, including their perspectives is warranted for future studies.

This study identified additional inpatient factors that impact appropriate testing in this setting, including out-of-pocket costs for diagnostic tests and delays in obtaining results. Access to quality laboratory services is a barrier to effective care in sub-Saharan Africa and continues to be a challenge.^1,38 Future studies should examine factors affecting quality of care in the inpatient setting for meningitis patients, particularly laboratory and supply chain barriers.

This study was conducted at a third-level referral hospital in Zambia. Therefore, we may not have adequately captured delays to meningitis care that differ in rural areas where access to biomedical care is limited and inpatient barriers to adequate meningitis management may be greater. Conversely, rural facilities may be less crowded than those in urban settings, and providers may be better positioned to explain disease etiology and treatment plans to patients and their caregivers in a common language. Further, nearly half of the interviews were conducted via telephone to facilitate social distancing for COVID-19. Transitioning from in-person to telephone interviews may have limited the candor of participants as it may have been more difficult to establish a rapport with patients and caregivers. However, the anonymity provided by a telephone interview may have allowed doctors and nurses to feel more open about sharing their challenges associated with caring for patients with meningitis.

Conclusions

While the burden of meningitis has improved, it continues to cause high mortality in regions with high HIV prevalence.^3,9 Patient pathways to adequate meningitis care are complex and influenced by delays in the home, at the clinic level and after hospital admission. In Zambia, improvements are needed in the care of patients with meningitis, including with the referral system, testing procedures, feedback of test results and initiation of treatment. People with conditions that put them at increased risk for severe neurologic disease, especially PLWH and their caregivers, should be counseled during routine clinical care regarding the clinical presentation of meningitis, causes and recommended course of action should symptoms occur.

Authors’ contributions

MAE, DS and PJW conceived the study; MAE, DS and PJW designed the study protocol; MAE, GM, MM, NN, MC and AMM conducted data collection; MAE, BC, GM, MM and NN conducted data analysis; GLB, MPK and IS aided with interpretation of results; MAE, BC and PJW drafted the manuscript; all the authors critically revised the manuscript for intellectual content. All the authors read and approved the final version of the manuscript. MAE is the guarantor of the paper.

Funding

This work was supported by the National Institutes of Health [grant numbers 5R25NS065729-11, 5R25NS089450]; the American Academy of Neurology Resident Research Scholarship; the American Neurological Association International Outreach Travel Scholarship; and the Johns Hopkins University Paul S. Lietman Global Health Travel Grant.

Competing interests

None declared.

Ethical approval

Ethical approval for this study was obtained from the University of Zambia's Biomedical Research Ethics Committee (UNZA BREC), the Zambian National Health Research Authority (NHRA) and the Johns Hopkins Medicine Institutional Review Board (JHM-IRB).

Data availability

Data are available upon reasonable request from the corresponding author.

References